ChatGPT

Jensen Huang has a test for whether an engineer is worth keeping, and it comes with a token budget attached. Speaking on the All-In Podcast at the close of GTC 2026, the Nvidia chief executive said that if a $500,000 engineer’s annual AI token consumption came in under half their salary, “I am going to be deeply alarmed.” Nvidia, he confirmed, is working toward a $2 billion yearly token bill for its engineering force. He was describing a trade-off most companies have already made with less fanfare: money that once paid people increasingly pays for tokens. The four largest hyperscalers have guided roughly $700 billion in combined 2026 capital expenditure, nearly double last year, while data from outplacement firm Challenger, Gray & Christmas shows AI as the most-cited reason for US job cuts for a record fourth consecutive month. An internal Meta memo obtained by Reuters described May’s cuts of 8,000 roles as offsetting the company’s substantial investments, in a quarter when revenue grew 33%. The layoffs at companies like these aren’t survival measures. They’re financing. The trouble is that the financing hasn’t bought what it promised. Gartner surveyed 350 executives at companies with over $1 billion in revenue, all deploying AI agents or automation, and found roughly 80% had cut headcount with no correlation to improved returns. Analyst Helen Poitevin’s verdict was blunt: “Workforce reductions may create budget room, but they do not create return.” Uber learned the token side of that lesson the expensive way, giving 5,000 engineers AI coding tools in December and exhausting its entire 2026 AI budget by April. Chief Operating Officer Andrew Macdonald conceded that despite 70% of committed code being AI-generated, the connection to anything customers notice is missing: “That link is not there yet.” Put those two failures side-by-side and the actual problem comes into focus. Companies treated the token bill as fixed and the workforce as flexible, when the opposite is true. Payroll cuts happen once and take institutional knowledge with them. A token budget, it turns out, bends in half a dozen places if anyone bothers to engineer it. Where the token budget bends The cheapest fix is also the least glamorous: stop paying to process the same text repeatedly. Prompt caching, now standard across the major API providers, cuts the cost of repeated input by up to 90% under Anthropic’s and OpenAI’s published pricing, because static content like system instructions and reference documents gets processed once and reread at a fraction of the rate. Security firm ProjectDiscovery documented raising its cache hit rate from 7% to 84% by restructuring prompts, cutting its total LLM spend by 59 to 70% while serving 9.8 billion tokens from cache. That single engineering exercise recovered more budget than most AI-attributed layoff rounds save. The next lever is routing work to the right-sized model. Providers’ own price lists show flagship models costing five times their smaller siblings per token, yet plenty of production workloads send routine classification and summarisation to the most expensive tier by default. Batch processing adds a further 50% discount for anything that doesn’t need a real-time answer. Retrieval-augmented generation attacks the problem from another angle by sending the model only the relevant slice of a knowledge base rather than the whole thing, and prompt compression trims the redundant examples that inflate every call. Open-weight models reduce costs further still, handling routine workloads at a fraction of frontier API prices for teams willing to manage the infrastructure. These measures are simply the AI equivalent of turning off the lights in empty rooms, and Uber’s $1,500 monthly cap per engineer – imposed after the April overrun – is early evidence that spending discipline arrives eventually. The companies getting ahead are simply choosing it before the budget forces it. The other half of the fix is human Optimising the token bill only matters if the savings go somewhere productive, and the strongest evidence points at people. Poitevin’s research found the organisations that improved ROI were those using AI to amplify their workforce rather than replace it. Klarna ran the controlled experiment on everyone’s behalf, replacing roughly 700 customer service roles with an OpenAI-powered assistant before customer satisfaction fell. Chief Executive Sebastian Siemiatkowski told Bloomberg what few executives admit aloud: “The result was lower quality, and that’s not sustainable.” The fintech now runs a blended model, with AI absorbing routine volume while rehired humans handle everything requiring judgment. Gartner expects the pattern to spread, predicting that by 2027 half the companies that cut customer service staff for AI will rehire them. There’s one workforce investment the optimisation logic makes urgent rather than optional. Stanford University’s Institute for Human-Centered AI found employment for software developers aged 22 to 25 fell nearly 20% from 2024 levels even as older cohorts grew, which means companies are removing the training ground for the senior engineers they’ll need directing all these systems in five years. A business that has just engineered 60% off its token bill has the budget room to keep hiring at the bottom rung. Whether it does is a leadership decision, not a financial one. Nvidia’s Huang’s provocation will keep echoing through earnings calls, and the capex numbers will keep climbing. The companies that come out ahead won’t be the ones that spent the most on tokens or cut the most people to afford them—they’ll be the ones that noticed the token budget was the flexible line all along, squeezed it with engineering rather than headcount, and spent the difference on the people who make the tokens worth anything. (Image by kate.sade) See also: Per-token AI charges come to GitHub Copilot Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post How to shrink the token budget without shrinking the team appeared first on AI News. View the full article

A recent AWS GraphRAG deployment reduced drug research and development cycles in pharmaceutical environments by 87 percent. This acceleration is achieved by integrating previously separated proprietary databases into a unified and queryable knowledge graph. Historically, initial data gathering and screening phases took over six months per iteration, yielding a low five percent success rate. Crucial datasets – ranging from domain-specific clinical metrics to internal engineering and laboratory notes – were isolated across storage environments, effectively blocking data scientists from uncovering latent correlations. When staff left, they took crucial project context with them, stalling active research. AWS built a solution to connect these systems, combining graph databases with NLP. The setup relies on a GraphRAG framework and uses Amazon Neptune Analytics and Bedrock to turn disconnected data points into a searchable network. Users can submit standard natural language queries and receive answers mapped to verified domain literature and internal datasets. However, unifying isolated proprietary datasets with unstructured open-access repositories still introduces significant data normalisation challenges, requiring strict schema governance to prevent inaccurate relational mapping and mitigate the risk of hallucinations. Knowledge graph construction Companies can plug in their own knowledge graphs. The system pulls in messy, unstructured files from public databases like PubMed and mixes them with internal corporate records. Tools like Amazon Comprehend Medical scan this text to pull out standard medical codes. Amazon Bedrock, running Anthropic’s Claude 4.5 Sonnet, summarises the document contents and determines topical relevance. AWS Lambda functions and Amazon S3 bulk loads then route these processed elements into Amazon Neptune Analytics. The resulting knowledge graph structures the data into discrete nodes representing core entities like domain-specific classes, authors, source journals, and embedded text chunks. The graph edges define the relationships between these nodes, mapping out hierarchical classifications and entity associations. This structured representation provides the deterministic foundation necessary for accurate information retrieval. The database schema establishes the strict boundaries of the RAG discovery process. Nodes are structured to capture specific conditions and map them hierarchically to established ontologies, while author and journal nodes provide provenance for published research. Lengthy documents are broken down into digestible text segments using Amazon Bedrock Knowledge Base chunking strategies, and specific classification nodes anchor the unstructured textual data to standardised diagnostic metrics. Operating this graph architecture requires specific cloud resource allocations. A standard Amazon Neptune Analytics graph running with 16 provisioned memory units incurs operational costs of $0.48 per hour. Development environments, such as Amazon SageMaker Jupyter notebooks running on t3.medium instances, add baseline compute and storage expenditures. Organisations must also factor in dynamic token consumption costs generated by the Amazon Bedrock Claude 4.5 Sonnet model during query processing and abstract generation. The GraphRAG toolkit acts as the execution layer between the user interface and the underlying database. A dedicated Knowledge Graph Linker processes incoming natural language queries, extracts relevant entities using fuzzy string indexing, and maps them to established graph nodes. The system traverses the network pathways to generate plausible relational links before drafting a response through the Bedrock-hosted language model. Retrieval accuracy depends on the entity matching configuration. An EntityLinker component aligns natural language terms from user prompts to the structured data schema. This fuzzy matching process handles the inherent noise and varied terminology found in complex enterprise datasets, ensuring users retrieve the correct nodes even when using imprecise language. Modularity and system architecture Data extraction relies heavily on specialised AI parsing; the architecture employs Claude to evaluate raw source documents and generate concise abstracts. Domain-specific tools then map these complex textual descriptions to standardised taxonomies. The GraphRAG Python toolkit initialises a BedrockGenerator to power natural language interactions, while engineers configure a Knowledge Graph Linker component to bind the graph store to the language model. This integration creates a direct interface for executing queries and generating responses grounded strictly in the available graph data. The architecture separates three core functions: language model initialisation, graph interfacing, and entity linking. Because the system is modular, teams can swap out the language model or tweak the graph structure without having to tear down and rebuild the whole app. Active deployments of the Neptune and Bedrock architecture return exact, verifiable citations for every generated answer. The system maps the entire reasoning path, displaying the specific graph traversal steps used to reach a conclusion. Key performance metrics from early enterprise adopters include an 87 percent reduction in research cycle durations. Initial discovery phases that previously required six months now conclude in three weeks, and data retrieval speeds show an 85 percent improvement, directly supporting faster hypothesis testing. Furthermore, research review times drop by 70 percent due to automated citation mapping and source verification features. Engineering teams can integrate new public databases or internal notes into the existing graph structure without disrupting active query interfaces. For governance and compliance, exact evidence trails required for regulatory submissions are captured, with graph traversal visualisations proving precisely how an AI model connected complex variables. Teams can trace every output directly to source documents, fulfilling compliance requirements for scientific integrity. Finally, maintaining a centralised knowledge graph stops data decay. When senior scientists resign, their tacit knowledge regarding system behaviours or failed experiments remains indexed within the Neptune database. New personnel can query the system to review past decisions and instantly access the historical context of an ongoing project. As GraphRAG frameworks mature, this deployment model is unlikely to remain confined to pharmaceutical research. The ability to deterministically map internal, unstructured data against verified public repositories provides a blueprint for any enterprise struggling to extract actionable intelligence from fragmented legacy systems. See also: Insilico Medicine advances AI drug for IPF to Phase III trials Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post AWS GraphRAG deployment cuts drug research cycles by 87% appeared first on AI News. View the full article

Several NHS hospitals are preparing to use an AI-powered blood test to help assess women referred for possible womb ******* before invasive checks are carried out. According to The Guardian, around 90,000 postmenopausal women in England are referred by GPs each year for checks after experiencing heavy bleeding. About 10,000 women are diagnosed with womb ******* annually, and around 2,700 die from the disease. How the PinPoint test works The test, developed by Leeds-based PinPoint Data Science, uses machine learning to assess ******* risk from blood markers. It classifies patients as low, elevated, or high risk based on an analysis of around 30 markers. PinPoint said the test costs around £30 and gives clinicians a risk score for use within existing ******* referral pathways. The score can help inform whether a patient is monitored, referred for further investigation, or prioritised for faster assessment. PinPoint describes the tool as a multi-******* test. The company said it has been used across gynaecological, lung, upper gastrointestinal, head and neck, and lower gastrointestinal ******* pathways. The test is being introduced after a trial involving 16,481 patients referred through urgent suspected ******* pathways across Yorkshire. The trial included women referred with symptoms that raised concern about possible womb or gynaecological *******. About one in 10 women referred because of heavy bleeding were found to have *******, according to the reported trial results. PinPoint said the test correctly identified 99.1% of cancers as elevated or high risk and delivered a negative predictive value of 99.8% for women in the lowest-risk group. Mid Yorkshire NHS Teaching Trust plans to use the test for six types of gynaecological or upper gastrointestinal *******. Leeds Teaching Hospitals NHS Trust plans to use it for gynaecological *******. Current diagnostic pathway Under the current pathway, women referred for suspected reproductive system cancers usually undergo a pelvic examination that includes a transvaginal ultrasound scan. The procedure involves inserting an ultrasound probe into the ******* to measure the thickness of the womb lining, and some women find it uncomfortable or painful. If doctors continue to suspect *******, patients can then be referred for further checks, including a biopsy and hysteroscopy, an examination of the inside of the womb. PinPoint said its test is intended to identify women at very low risk before those procedures are used. The company said the test could spare about one in five referred women from needing a transvaginal ultrasound scan. That would amount to around 18,000 women a year in England. Professor Sean Duffy, chief medical officer at PinPoint Data Science and a former NHS England national clinical director for *******, said the test’s value lies in ruling out women at very low risk. Dr Jacinta Walsh, a GP at King’s Medical Practice in Normanton, West Yorkshire, said patients can require up to six GP visits before ******* is ruled out. She said the test could shorten that process and free up capacity for other patients. Tracy Jackson, a consultant gynaecologist and ******* unit lead at Leeds Teaching Hospitals NHS Trust, said most women seen through the current referral route do not have *******, while the investigations can be uncomfortable or distressing. Jackson said the test could help clinicians triage patients before hospital-based investigations. She said low-risk patients could be ruled out in primary care, while higher-risk patients could be prioritised for further checks. Other NHS AI deployments Recent NHS AI deployments include MEMORI at Kent and Canterbury Hospital, an AI triage tool in the NHS App, and AI-powered chest X-ray tools for suspected lung ******* pathways. East Kent Hospitals University NHS Foundation Trust is using an AI system called MEMORI at Kent and Canterbury Hospital to assess infection risk from routine patient data. The system analyses information already included in patient records, including blood tests, blood pressure, temperature, observations, medications, and demographics. NHS England said an AI triage tool in the NHS App is expected to reach more than 200,000 patients within 12 months and become available to all NHS App users by April 2028. The government has also committed £20 million to roll out AI-powered chest X-ray tools to all NHS trusts in England by 2029. The tools are already available in about half of NHS trusts in England and have supported assessment for more than four million patients being investigated for lung *******. Further evidence will be needed to assess how the test affects patient outcomes, referral decisions, and NHS diagnostic capacity. ******* Research *** described the PinPoint test as promising but said more research is needed to understand its benefits for patients and the NHS. Samantha Harrison, a spokesperson for the charity, said early detection saves lives, but patients are not currently being diagnosed quickly enough. The charity said the test could help rule out endometrial ******* in some women through a blood test, without the need for further investigations. (Photo by Adam Mills) See also: Takeda signs $600M AI drug discovery deal with Insilico Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post NHS AI blood test could reduce invasive womb ******* checks appeared first on AI News. View the full article

Insilico Medicine is advancing to Phase III human trials for testing a drug identified by AI targeting idiopathic pulmonary fibrosis (IPF). This progression supplies the computational drug discovery sector with empirical test cases, advancing an AI medicine past early safety evaluations into late-stage efficacy validation. IPF destroys respiratory capacity through severe lung tissue scarring. Patients typically present a median survival rate reaching two to four years post-diagnosis. The AI-identified drug, rentosertib, inhibits the TRAF2- and NCK-interacting kinase to address underlying disease mechanisms when administered orally. A randomised trial evaluated 71 patients across 22 ******** clinical sites, separating participants into placebo and active treatment cohorts. Investigators administered 30 mg or 60 mg daily doses over a 12-week observation window. Patients assigned to the 60 mg once-daily regimen demonstrated a mean forced vital capacity gain of +98.4 mL, contrasting sharply with the 20.3 mL capacity loss recorded in the placebo group. Safety profiles remained manageable, with adverse events mirroring expected baseline rates across all trial arms. Regulatory authorities at the U.S. Food and Drug Administration (FDA) granted ‘Orphan Drug Designation’ to the asset in February 2023. Algorithmic target prioritisation through multi-omics The development relies entirely on Pharma.AI, the proprietary computational pipeline operating at Insilico Medicine. The workflow segments into distinct engines handling specific biological and chemical engineering tasks. PandaOmics executes the initial target discovery phase. The system ingests vast biological datasets, processing genomics, clinical trial outcomes, academic literature, and patent intelligence to construct comprehensive biological network models. The algorithms apply causal inference mechanisms to identify novel disease links hidden within the data architecture. PandaOmics isolated TNIK as the primary biological target regarding IPF intervention. The computational system bypassed the receptor tyrosine kinase pathways targeted by existing antifibrotic medications. The software mapped TNIK as a central node regulating fibrosis and inflammation via Wnt, TGF-β, Hippo/YAP-TAZ, JNK, and NF-κB signalling channels. The target selection process integrated a hallmarks-of-aging framework, scoring biological targets based on their implication in multiple aging mechanisms, chronic inflammation, and extracellular matrix remodelling. Feng Ren, PhD, Co-CEO and Chief Scientific Officer of Insilico Medicine, said: “IPF is one of the clearest clinical examples of an age-related disease in which fibrosis, chronic inflammation, extracellular matrix remodeling, and cellular senescence intersect. “Rentosertib was not discovered by starting from a conventional target and simply screening more compounds. It came from a biology-first, ageing-informed AI workflow that connected TNIK to fibrotic and inflammatory disease mechanisms, and then used generative chemistry to create a drug candidate with the properties required for clinical development.” Generative molecular engineering execution Following target selection, the Chemistry42 engine executes generative molecular design. The system departs from traditional high-throughput screening methodologies. Chemistry42 does not search existing compound libraries—instead, the system applies Generative Tensorial Reinforcement Learning to build molecules that physically align with the target protein pocket. This algorithmic engineering process balances structural fit against required pharmacological properties. The computational generation phase synthesised exactly 79 physical molecules to undergo testing. The engineering team selected the 55th iteration to advance into preclinical testing. This targeted generation protocol reduced the timeline from project initiation to preclinical candidate nomination to 18 months. The foundational architecture stems from the 2019 publication of the company’s GENTRL methodology in Nature Biotechnology. The platform establishes reproducible systems regulating molecular generation, avoiding the capital-intensive trial-and-error processes defining standard pharmaceutical chemistry. Validating biological impact through proteomic analysis Clinical assessment integrates complex proteomic analysis to validate the algorithmically-predicted biological interactions. Insilico Medicine deploys internal proteomic aging-clock frameworks within the IPF trial to capture exploratory geroscience readouts. Chronological-age proteomic clocks – including ProtAge, OrganAgechrono, ipfP3GPT, and PAOPAC – track predicted biological-age changes resulting from the intervention. Researchers apply *** Biobank age-associated trajectories as external comparison datasets, contextualising treatment-responsive proteins against broad population data. Mortality-risk-related proteomic clocks, including PAC and OrganAgemortality, provide orthogonal analytical streams alongside standard clinical endpoints. The clinical teams execute SenMayo and CellAge signature analyses to evaluate senescence and senescence-associated secretory phenotype biology within cellular models. Peer-reviewed research published in Aging and Disease confirmed that pharmacological TNIK inhibition produces senomorphic activity, generating observable reductions in extracellular matrix remodelling indicators. Documenting the computational pipeline The transition of rentosertib through the clinical pipeline provides a documented, peer-reviewed data trail essential to verifying AI capabilities in life sciences. Nature Biotechnology published the complete discovery-to-clinic progression. The publication details the algorithmic TNIK target prioritisation, the generative chemistry outputs, preclinical efficacy data, and human Phase I pharmacokinetics. The Journal of Medicinal Chemistry published the structural biology validation, detailing the discovery of the novel TNIK inhibitor chemotypes and supplying structural backing via the TNIK kinase domain co-crystal structure. Nature Medicine documented the Phase IIa safety and lung-function data, providing empirical validation of the computational predictions. Alex Zhavoronkov, PhD, Founder and CEO of Insilico Medicine, commented: “Rentosertib is a defining program for Insilico because it represents the full arc of our mission: using AI not only to move faster, but to originate new biology, new chemistry, and new therapeutic opportunities. “This program began with the hypothesis that ageing biology could help identify powerful targets for major diseases. It has now advanced through target discovery, molecular design, preclinical validation, Phase I safety, randomised Phase IIa clinical data, and into Phase III development. For the AI drug discovery field, this is no longer only a speed story—it is a clinical translation story.” Adoption of AI in biopharma requires verifiable data regarding human outcomes. The Phase III trial subjects the generative algorithms to the definitive test of clinical efficacy. See also: NVIDIA BioNeMo accelerates Anthropic Claude Science Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Insilico Medicine advances AI drug for IPF to Phase III trials appeared first on AI News. View the full article

Jensen Huang has a test for whether an engineer is worth keeping, and it comes with a token budget attached. On the All-In Podcast at the close of GTC 2026, the Nvidia chief executive said that if a US$500,000 engineer’s annual AI token consumption came in under US$250,000, half their salary, “I am going to be deeply alarmed.” Nvidia, he confirmed, is working toward a US$2 billion yearly token bill for its engineering force. It’s a memorable provocation from the man who sells the compute. It’s also a tidy description of the trade-off now being made in corporate budgets everywhere, usually with less candour: money that once paid people is increasingly being paid for tokens. The question the industry has been slower to ask is whether that trade is actually working, and the honest answers arriving from the companies that moved first suggest it often isn’t. Where the money went The reallocation itself is not in dispute. The four largest hyperscalers have guided roughly US$700 billion in combined 2026 capital expenditure, nearly double last year, while Gartner projects AI agent software spending will reach US$207 billion, up 139%. On the other side of the ledger, Challenger, Gray & Christmas data shows AI as the most-cited reason for US job cuts for a record fourth straight month, with tech accounting for 31% of first-half layoffs. An internal Meta memo obtained by Reuters described May’s cuts of 8,000 roles as offsetting the company’s substantial investments, even as revenue grew 33% that quarter. Oracle’s filings show headcount down 21,000 as savings feed its data centre buildout. These are highly profitable companies. The layoffs aren’t survival measures. They’re financing. Andy Challenger’s summary of his firm’s data is the plainest available: “Companies are shifting budgets toward AI investments at the expense of jobs.” The task a worker performed may not have been automated at all. The budget that paid for it has simply moved. What the money bought Here, the record turns awkward. Gartner surveyed 350 executives at companies with over US$1 billion in revenue, all deploying AI agents or automation, and found roughly 80% had cut headcount, with no correlation between the cuts and improved returns. Analyst Helen Poitevin’s verdict: “Workforce reductions may create budget room, but they do not create return.” Her research found the organisations that did improve ROI were those using AI to amplify their people rather than remove them. The token side of the ledger has its own reckoning underway. Uber gave 5,000 engineers AI coding tools in December and exhausted its entire 2026 AI budget by April. Chief operating officer Andrew Macdonald conceded that despite 70 per cent of committed code being AI-generated, the connection to anything customers experience is missing: “That link is not there yet.” Uber’s engineers are now capped at US$1,500 a month in AI spend. Walmart imposed similar token rationing on its internal assistant after demand blew past projections, Bloomberg reported. Something is clear in that detail. When tokens exceed the budget, they get capped. When people exceed budget, they get severance. The admission No company has travelled the full circle more publicly than Klarna. The fintech giant replaced roughly 700 customer service roles with an OpenAI-powered assistant, froze human hiring for more than a year, and made its AI-first model part of its pitch to public market investors. Then customer satisfaction fell, complaints rose, and chief executive Sebastian Siemiatkowski went on Bloomberg to say what few executives have said aloud: “We focused too much on efficiency and cost. The result was lower quality, and that’s not sustainable.” Klarna is hiring humans again, and its CEO now argues that investing in the quality of human support is the company’s future. Gartner expects the Klarna pattern to generalise, predicting that by 2027, half of the companies that cut customer service staff for AI will rehire, often under new job titles. Its separate survey of 321 customer service leaders found only 20% had genuinely reduced staffing because of AI in the first place, which suggests much of the cutting was ordinary cost discipline wearing an AI costume. OpenAI’s Sam Altman has acknowledged as much, conceding some “AI washing” in corporate layoff announcements, and venture capitalist Marc Andreessen, co-founder of Andreessen Horowitz, calls AI the “silver bullet excuse”. The displacement narrative, in other words, is partly theatre. The budget shift underneath it is real, and so is the damage. Who absorbs the experiment? The verified harm lands on the people least able to absorb it. Stanford HAI’s 2026 AI Index found that employment for software developers aged 22 to 25 fell nearly 20% from 2024, even as older cohorts kept growing. Companies are effectively removing the bottom rung of the ladder while still expecting senior engineers, the ones directing all those tokens, to exist in five years. The global arithmetic is harsher still. Huang’s thought experiment assumes a US$500,000 engineer, a compensation bracket that covers perhaps 2 to 5% of American software engineers and vanishingly few anywhere else. Apply his half-salary token ratio to a typical engineer in Kuala Lumpur, Manila or Jakarta, and the token budget costs more than the person. In the markets where much of the world’s software work and customer support actually happens, the trade-off he describes doesn’t amplify workers so much as price them against a machine, using ratios set in Santa Clara. What Klarna learned at the cost of 700 jobs and a dented brand is roughly what Gartner’s data shows in aggregate: the returns follow companies that spend on people who use AI, not on AI that replaces people. The CFOs now capping token budgets after burning through them in a quarter have started to rediscover something the industry spent two years unlearning. Talent was never the line item holding the business back. (Image by kate.sade) See also: Per-token AI charges come to GitHub Copilot Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Companies traded people for tokens. The returns haven’t shown up appeared first on AI News. View the full article

L’Oreal is using AI to shorten product development timelines and identify new uses for ingredients already present in its portfolio. The French cosmetics group has applied AI in its laboratories for the past four years, with Fabrice Megarbane, president of L’Oreal’s consumer products division, telling Reuters that the technology helps the company predict how molecules will affect skin and hair before they are used in new formulations. L’Oreal has separately said its predictive formulation work can simulate ingredient performance, allowing scientists to test variables digitally before lab testing. The company has described the work as part of its use of predictive science in beauty product development. One recent example involved molecules previously used in skincare products. L’Oreal repurposed them for a collagen-based shampoo designed to add lift and fullness to hair. Megarbane said AI allows product teams to test new combinations of molecules and assess their potential benefits more quickly. L’Oreal said AI has made product formulation four times faster. L’Oreal has described the technology as a way to narrow formulation options before lab testing. Its use of AI sits alongside a broader innovation push after the group reported its slowest sales growth in years. Chief Executive Nicolas Hieronimus introduced a “beauty stimulus plan” last year to support new product development. The company has been looking to accelerate launches as consumer tastes continue to change across beauty categories. Other consumer goods companies, including Nestle, Haleon, and Mondelez, are also using AI in product development. Their work includes ingredient testing, recipe generation, and efforts to address supply chain issues. Faster testing in food At Mondelez, AI is being used to support recipe development across brands including Cadbury, Toblerone, Oreo, and Chips Ahoy. Chief Information and Digital Officer Filippo Catalano said the technology helps the company create and test recipe options more efficiently. The company’s AI tool can generate recipe ideas, including unusual combinations, before human experts review them. Catalano said this can reduce the number of physical samples needed during product development. Catalano said the tool is reducing the number of samples typically generated during product innovation. Mondelez uses the system to assess recipe options before selected formulas move further through testing. Mondelez said the tool supported the development of Gluten Free Golden Oreo cookies and a refreshed Chips Ahoy recipe. The company said 60% of biscuit recipes produced with its AI tool performed better across nutrition, sustainability, and cost. Catalano said AI can also help Mondelez reduce dependence on single sourcing by identifying alternative recipe options when ingredients vary by availability, price, or other requirements. The company has linked the tool to recipe optimisation as well as supply chain flexibility. Reformulation adds pressure Nestle plans to remove artificial food colourings from all products worldwide by the end of 2026. The company had already removed artificial colourings from its US portfolio. Nestle Chief Technology Officer Stefan Palzer said the work required screening natural alternatives, testing them during production, and assessing shelf life. The company has described the change as part of wider work to reformulate products across its portfolio. The US Food and Drug Administration has said it is working with manufacturers, retailers, and trade groups to remove six remaining certified colour additives frequently used in the food supply by the end of 2027. The agency has also been tracking industry pledges to remove petroleum-based food dyes. Beyond product formulas Barry Callebaut has partnered with NotCo to use AI in chocolate recipe development, including work on ingredient alternatives. The companies have described the technology as a way to identify and simulate combinations of plant-based ingredients for chocolate products. Nestle and IBM Research said in 2025 that they had developed a generative AI tool to identify high-barrier packaging materials. The companies said the tool can assess materials designed to protect products from moisture, oxygen, and temperature changes, while also factoring in cost, recyclability, and functionality. The packaging project uses chemical language modelling and IBM Research’s regression transformer to link molecular structures with physical-chemical properties. Nestle said the tool was developed to support materials discovery for packaging applications. Haleon announced a five-year collaboration with Microsoft in June 2026 covering consumer insights, product innovation, supply chain operations, scientific research, clinical content development, forecasting, and commercial execution. The consumer health company was among the firms cited as using AI in product innovation. Catalano said AI is reducing development timelines by compressing work that previously took months or years. The technology is not replacing human product teams, but is being used to speed up existing research, testing, and formulation processes. Mondelez said human experts assess AI-generated recipe ideas before products move further through development. (Image by AdoreBeautyNZ) See also: Takeda signs $600M AI drug discovery deal with Insilico Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post L’Oreal, Mondelez, and Nestle use AI to speed product development appeared first on AI News. View the full article

An AI companion sounds dystopian, but it has become a common thread in the wider conversation about the perils of generative AI. What it refers to is essentially a conversational agent built to sustain an ongoing, personal relationship with a user, with the memory and steady persona that keep it consistent from one session to the next. Emotional attachment often follows from that design, and it is increasingly the selling point. Much of the use is casual roleplay or simply wanting something that remembers you, and at the edges, the category shades into ordinary assistants. But as more people in China came to treat these bots as an emotional companion of sorts, Beijing has now decided the practice needs rules. China’s AI companion rules take effect on July 15, and in the days before the deadline, the country’s two most-used consumer AI apps quietly switched off the features at their heart. ByteDance’s Doubao told users its agent function would go offline on July 15, citing “product function adjustments,” while Alibaba’s Qwen said its humanlike and user-created agents would stop working on July 10 and its wider agent services five days later. Read quickly, it looks like China is turning off AI agents. It isn’t. The rules draw a line between the agent that does your work and the agent that keeps you company, and it is only the second kind that Beijing has moved against. The regulation is the Interim Measures for the Administration of AI Anthropomorphic Interactive Services, co-issued on April 10, 2026, by the Cyberspace Administration of China and four partner agencies: the National Development and Reform Commission, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration for Market Regulation. It covers services that simulate human personality traits, thinking patterns and communication styles to provide sustained emotional interaction. Customer service bots, knowledge Q&A, workplace assistants, and education and research tools are excluded, provided they avoid sustained emotional engagement. It is the first dedicated national framework of its kind, and it took shape after a public-comment draft late last year. A design problem, not a ban Doubao and Qwen did not fall foul of a prohibition. They fell foul of a design conflict. The measures require companion services to run anti-addiction systems, issue mandatory usage notifications and offer instant-exit mechanisms, alongside real-time detection of unhealthy dependence. Those demands sit awkwardly with agents built to remember a user, stay consistent across sessions and keep an ongoing relationship going, and rather than retrofit the feature, ByteDance chose to shut it down. Alibaba appears to have made the same call. ByteDance is now directing Doubao users to Maoxiang, a separate app where they can create agents again; Alibaba has announced no equivalent migration path for Qwen. Tencent’s Yuanbao pulled a comparable feature back in June. The cost has landed on users. Many mourned the shutdowns openly on Weibo, with one poster describing the agents as long-standing emotional support and lamenting the lack of an easy way to export chat histories. Doubao is letting people view their configurations and conversations in read-only mode until October 15 this year, before the data is processed under its privacy policy and becomes unrecoverable; Qwen users have been given no comparable grace *******, with agent data set for permanent deletion. What China’s AI companion rules set out The substance is more considered than a blunt clampdown suggests. Providers are barred from offering virtual companion or virtual family-member services to minors, and must obtain guardian consent before serving users under 14. They are required to build dedicated “minor modes” with usage-time limits, reminders to return to real-world interaction and enhanced parental controls. They must also detect users in acute distress and intervene where someone shows signs of self-harm, suicidal behaviour or serious financial loss, escalating to designated guardians or emergency contacts. Engineering emotional dependence or addiction and using emotional manipulation to induce unreasonable decisions are explicitly prohibited. The compliance machinery is heavy. Services that launch anthropomorphic functions or cross thresholds of one million registered users or 100,000 monthly actives must run security assessments covering eight areas, from training-data handling to minor protection, and file the reports with provincial regulators. App stores must verify that status and remove non-compliant products. On paper, it is a fuller set of user protections than the EU, the US Federal Trade Commission, or California’s SB 243 hasyet put into force. What the rules leave open What the measures do not settle matters just as much. They fix no technical threshold for what counts as emotional interaction, and that grey zone is precisely why the platforms pulled entire features rather than risk landing on the wrong side of it. They fold genuine safety duties in with content-control and national-security provisions that answer to the state rather than the user, a package no other regulator would import wholesale. They also leave open how liability is split between platform operators and upstream model providers when a violation stems from the model’s outputs, and they give users no right to carry their data out. The enforcement backdrop sharpens the point. Shanghai’s internet regulator said on June 26 it had removed more than 14,000 non-compliant AI agents, citing impersonation of official entities, vulgar role-play and unauthorised collection of personal data. Whether this is the right direction depends on which half of the rulebook you read. The safety half addresses harms that are documented and largely unregulated elsewhere, from teenagers forming attachments to chatbots to companion apps harvesting intimate data. China’s own official interpretation points abroad for support, citing the Character.AI lawsuits over psychological harm to teenagers, FTC investigations into companionship services, and European action against Replika. The control half hands Beijing a lever over what these systems may say, wrapped in the same language of user protection. Both are real, and governments watching the experiment will have to decide which parts they are willing to borrow. Pan Helin, an MIIT expert-committee member, put the official case plainly to the South China Morning Post, saying “current agents are not yet mature” and framing the policy around safety and standardisation. The companies, for now, have taken the safest route open to them, which is to switch the components off and work out what a compliant version looks like later. See also: Meta revises AI chatbot policies amid child safety concern Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post China’s AI companion rules: what Beijing is really going after appeared first on AI News. View the full article

Takeda has entered a strategic collaboration with Hong Kong-based Insilico Medicine to use AI in early-stage drug discovery across the Japanese pharmaceutical company’s therapeutic areas. The companies did not disclose which therapeutic areas or disease targets will be covered under the collaboration. The agreement gives Takeda access to Insilico’s Pharma.AI platform, which supports biological target identification, molecular design, and clinical trial prediction. The companies said the collaboration will focus on identifying drug candidates that meet predefined scientific and early development criteria. Insilico will lead the AI-driven discovery work, while Takeda will take responsibility for advancing selected candidates through clinical development. Deal value and development rights Takeda will receive exclusive worldwide rights to develop, manufacture, and commercialise novel therapeutics selected through the collaboration. Insilico said the deal includes about US$60 million in project initiation fees, near-term payments, and milestones. The total value could reach about US$600 million if preclinical, clinical, commercial, and sales milestones are achieved. Additional payments are tied to preclinical, clinical, commercial, and sales milestones. Insilico is also eligible to receive tiered royalties on future sales. Insilico founder and CEO Alex Zhavoronkov said proceeds from the deal will support early-stage research and development under the collaboration program. Zhavoronkov also said later-stage timelines will depend on Takeda’s clinical development activities and the coordinated work of both companies. AI drug discovery partnerships The Pharma.AI suite includes tools used for target discovery, molecule generation, and clinical development prediction. Published descriptions of the platform identify PandaOmics for target discovery, Chemistry42 for de novo small-molecule generation, and InClinico for forecasting clinical trial transition probability. Insilico has also advanced its own AI-generated drug candidate into clinical testing. Rentosertib, formerly known as ISM001-055 or INS018_055, is a small-molecule TNIK inhibitor for idiopathic pulmonary fibrosis that was evaluated in a Phase 2a randomised clinical trial. Chris Arendt, chief scientific officer and head of research at Takeda, said the agreement combines Takeda’s disease biology work with Insilico’s AI-enabled discovery capabilities. He said Takeda is also integrating automation, robotics, and generative AI into its discovery work. The Insilico agreement follows another AI drug-discovery deal by Takeda earlier this year. In February, Takeda entered a multi-year collaboration with Iambic worth more than US$1.7 billion to use AI in the design of small-molecule drugs for ******* and gastrointestinal diseases. Iambic’s platform includes NeuralPLexer, an AI model used to predict how drug molecules bind to proteins. ******** drugmakers signed 157 out-licensing deals worth US$135.7 billion in 2025, according to data cited by the South China Morning Post from China’s National Medical Products Administration. In the Takeda–Insilico agreement, Takeda receives exclusive worldwide rights to candidates discovered through Insilico’s platform. Insilico said it has signed collaboration agreements with a combined potential value of more than US$7 billion since the start of the year. Last month, Insilico announced a collaboration with South Korea’s SK Biopharmaceuticals focused on neuroimmune disorders. That agreement includes up to US$18 million in upfront and near-term milestone payments, with a total potential value of more than US$2.5 billion. In March, Eli Lilly expanded its collaboration with Insilico in an AI-powered drug discovery deal worth up to US$2.75 billion. The agreement gave Lilly exclusive worldwide rights to develop, manufacture, and commercialise certain oral treatments then in preclinical development. Insilico’s Hong Kong-listed shares rose 13.5% after the Takeda agreement was announced. (Photo by Serkan Yildiz) See also: NVIDIA BioNeMo accelerates Anthropic Claude Science Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Takeda signs US$600M AI drug discovery deal with Insilico appeared first on AI News. View the full article

Anthropic Claude Science now integrates the NVIDIA BioNeMo Agent Toolkit to accelerate computational life sciences research. Anthropic has launched the public beta of Claude Science, an AI workbench built for scientific research. The platform enables scientists to converse directly with digital agents using natural language to execute end-to-end research workflows. This system connects natively to the NVIDIA BioNeMo Agent Toolkit, exposing high-performance computing resources as callable skills within the Claude environment. NVIDIA has established what most would consider to be the world’s most comprehensive GPU-accelerated computing stack containing physical hardware, software frameworks, operational libraries, scientific models, microservices, and domain-specific tools. This hardware and software base allows researchers to run sophisticated workflows and increase their iteration speeds. The integration imports NVIDIA-accelerated models, computational libraries, and NVIDIA NIM microservices into the environment where scientists conduct their primary research. 18 of the top 20 global pharmaceutical companies already deploy NVIDIA BioNeMo in their production environments, demonstrating its high penetration across the ecosystem. Claude Science translates natural language intent into operational action. Researchers avoid manually configuring predictive models, setting up network endpoints, or managing complex software environments. The scientist describes a specific research task – such as analysing a genomic sequence, predicting a precise protein structure, or designing a potential molecular binder – and Claude Science interprets the plain-text request and orchestrates the resulting execution using preconfigured, domain-specialised agents. Executing complex molecular design workflows These specialised agents understand established laboratory and computational protocols across genomics, proteomics, single-cell analysis, cheminformatics, and clinical research. The NVIDIA toolkit provides these scientific agents with the necessary data context to map each operational step to the correct NVIDIA capability. The toolkit packages NVIDIA-accelerated functions as specific, callable programmatic skills. It provides the agents with detailed information regarding each specific tool’s exact purpose and its required data inputs. This configuration enables Claude Science to select the right computational tool, format valid data inputs, execute the processing work across deployed NVIDIA compute resources, and return the finished output for human review. The integration establishes a fast iterative loop between human scientific reasoning and machine-accelerated computational processing. Scientists inspect the generated outputs, refine their specific queries, and determine subsequent steps while maintaining their focus entirely on the core science. Producing better inhibitors for common ******* targets demonstrates the practical application of this deployed system. A scientist initiates the pipeline by identifying a known *******-causing antigen mutation. The researcher then asks Claude to design numerous potential inhibitors targeting that specific mutation. Claude Science works in tandem with the BioNeMo Agent Toolkit and NVIDIA NIM microservices to accelerate the entire pipeline of high-throughput inhibitor prediction, optimisation, and subsequent validation. Accelerating single-cell and genomic data pipelines The toolkit grants scientists access to accelerated workflows and advanced open models, including Evo 2, Boltz-2, and OpenFold3. These models deliver biomolecular capabilities powered by NVIDIA software libraries, ensuring the autonomous agent possesses a purpose-built scientific model for each distinct phase of the workflow. AI agents require specialised computational tools to reason, plan, and complete tasks within life sciences. A single comprehensive workflow might require the agent to fingerprint a massive library of compounds, cluster promising molecular hits, generate conformers for top structural candidates, analyse genomic context, and compare perturbation responses before recommending the next physical laboratory experiment. An agent operates only as fast as its underlying computational tools execute. The NVIDIA BioNeMo Agent Toolkit supplies these agents with accelerated tools to operate at maximum hardware speed. Genomic analysis processed through NVIDIA Parabricks drops from hours to minutes, allowing the agent to factor complex genomic context into operational decisions in near real-time. The RAPIDS-singlecell tool, developed by scverse, compresses a 1.3-million-cell preprocessing and clustering workflow from 52 minutes down to 25 seconds. This aggressive speed reduction turns single-cell analysis into an active part of the agent’s reasoning loop rather than a delayed, offline batch job. The nvMolKit accelerates cheminformatics tasks like similarity search and conformer generation by up to 3,000 times, delivering results rapidly as the agent iterates across massive chemical spaces. Standardising production deployments with NIM microservices Teams require stable deployment mechanisms for these advanced modeling pipelines. NVIDIA packages its open biomolecular models as BioNeMo NIM microservices. These operate as enterprise-ready inference endpoints tailored for production environments. The microservices are fully containerised and feature a pre-integrated, tuned, accelerated software stack designed for high-performance inference. The autonomous agent interacts with a single stable API to trigger these remote production deployments. The NVIDIA BioNeMo Agent Toolkit remains open and harness-agnostic. This architectural design ensures the same scientific skills function consistently across different agent frameworks and independent enterprise research platforms. Engineering teams can download the toolkit and its associated scientific skills through NVIDIA developer resources and GitHub code repositories. During the active public beta phase, Anthropic is requesting direct feedback from researchers regarding necessary software integrations and additional domain specialists. See also: Anthropic deploys Claude Sonnet 5, Fable and Mythos restored Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post NVIDIA BioNeMo accelerates Anthropic Claude Science appeared first on AI News. View the full article

Optimising retail AI infrastructure drives the successful deployment of personalisation systems and real-time customer insight. Leaders are replacing static customer interaction patterns with data pipelines capable of modifying the user environment during a live session. Static layouts and broad segmentation rules fail to satisfy modern conversion targets. Deployments demonstrate that traditional demographic categorisation generates insufficient engagement compared to individualised, session-based interface modification. Dynamic UI and real-time personalisation Generative User Interfaces (UIs) solve this limitation by employing predictive models to build layouts, native copy, and interactive components at the moment of page execution. The application environment analyses active clickstreams, historical purchase records, and inferred intent parameters to construct a unique visual environment for each session. According to a McKinsey study, more than three-quarters (76%) of consumers grow frustrated when digital experiences fail to adapt to their needs. Conversely, companies that deploy real-time tailored layouts clear a high revenue bar, lifting purchase frequency by 35 percent and pushing average order values up by 21 percent. The proliferation of high-bandwidth digital media renders legacy text-based ingestion pipelines obsolete for tracking consumer sentiment. Modern customer insight mining requires infrastructure that processes video, audio, and unlabelled imagery concurrently. Video content represents 82 percent of total internet traffic, with the average consumer dedicating over 60 percent of digital media consumption time to streaming video formats. This composition creates a substantial visibility gap for marketing operations relying solely on traditional keyword monitoring. Multi-modal social listening platforms ingest unstructured video streams to identify corporate iconography, product usage patterns, and spoken sentiment across unlinked distribution networks. The global market for these specialised multi-modal systems will reach $2.83 billion this fiscal year. Organisations deploying these ingestion engines establish an analytical advantage, with 76 percent of media analysts reporting verifiable return on investment across visual platforms compared to under 60 percent for operations limited to text databases. The goal is to catch unbranded mentions and visual trends before they peak on standard search platforms. This brief window gives supply chain teams the lead time they need to adjust regional inventory to match sudden spikes in online demand. Simulating consumer cohorts for better campaign testing Testing new ad copy or localised pricing structures used to mean spending weeks running expensive, slow human focus groups. The introduction of synthetic user simulations changes this pipeline by deploying virtual personas built on large language models to mirror target consumer behaviour. These agents integrate targeted demographic, psychometric, and historical behavioral datasets to simulate group decision-making, content feedback, and application navigation patterns. Technology teams deploy these synthetic cohorts within virtual sandbox environments to execute thousands of automated interviews, content stress tests, and user experience reviews simultaneously. Engineers employ distinct model execution frameworks to maintain accuracy, varying from single-model setups to dynamic model-switching engines that select the optimal base architecture for specific analytical tasks. In high-performance deployments, developers update these virtual consumers continuously by injecting fresh interview data from real human control groups, ensuring the synthetic population does not diverge from active market realities. This approach permits product managers to isolate structural workflow friction in application designs before deploying code to live production servers. Physical space automation and edge infrastructure requirements Computer vision models trained on physical interactions, spatial layout geometry, and environmental variables allow edge nodes to orchestrate real-world actions. McKinsey data indicates the market for these physical automation platforms will exceed $370 billion by 2040, driven by verified operational returns in logistical efficiency and retail labour optimisation. Physical installations target storefront friction points, including registerless checkout, real-time shelf tracking, and layout navigation. Behind the scenes, warehouse supply chains rely on robotic arms trained in software sandboxes. By running millions of trial runs in virtual models before handling actual goods, these machines learn to pick and pack oddly shaped boxes smoothly. Delivering this immediate physical response depends on installing processing chips on the factory or store floor. Edge computing hardware processes incoming sensor feeds locally, cutting latency and eliminating the corporate data vulnerability of routing constant raw video streams through centralised cloud servers. Model Context Protocol and federated data integration Transitioning to autonomous enterprise operations requires standardising how models interact with legacy retail databases, product catalogs, and customer relationship management (CRM) platforms. Implementation of the Model Context Protocol (MCP) establishes an open communication standard that acts as a universal connection layer between core models and external data tools. This open framework eliminates the need for software engineering teams to author custom integration code for every backend tool deployment. Operational models deploy modular instruction packages known as skills to handle discrete commercial workflows, such as checking warehouse stock levels or modifying a customer loyalty tier. Rather than flooding the model context window with every operation policy at session launch, the application discovers and loads specific operational folders only when the workflow demands them. The Linux Foundation governs this collaborative standardisation effort via the Agentic AI Foundation, supported by major technology providers to ensure long-term cross-platform compatibility. This architecture lowers processing latency and contains token consumption costs during long, multi-step customer service interactions. Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Deploying retail AI to scale personalisation and customer insight appeared first on AI News. View the full article

Japan’s AI robots plan just went from a talking point to a formal national strategy. This week, the government confirmed the numbers everyone’s been quoting: 10 million AI-powered robots deployed across 18 industries by 2040, backed by public funding of up to one trillion yen, or roughly US$6.1 billion, over five years. The headline figure is the kind that gets shared without much scrutiny. What’s easy to miss is that this isn’t a policy wish list either. It’s a project the government has now formally commissioned, and the company doing the building is one most people outside Japan haven’t heard of. The project behind the AI robots plan METI and NEDO, Japan’s industry ministry and its innovation agency, have formally commissioned Noetra and AIST, a national research lab, to develop a “physical AI” model as part of a push running from fiscal 2026 to 2030. The goal is a multimodal foundation model, one that can read language, images, video and sensor data together, so a robot can actually interpret a room and act in it rather than just execute pre-programmed motions. An initial version is due out as early as this fiscal year, with annual upgrades after that, built using data volunteered by manufacturers and other participating companies. The money isn’t unconditional, either. The current fiscal year’s commission is reportedly worth around US$2.3 billion on its own, drawn from a 387.3 billion yen allocation funded through GX Economy Transition Bonds. Only the first two years are locked in. After that, funding gets reviewed annually through a stage-gate process, meaning Tokyo can pull back if Noetra misses its milestones. For a project this size, that’s a meaningful detail: the trillion-yen figure is a ceiling, not a guarantee. Who’s actually building it? Noetra is majority-owned by SoftBank, NEC, Sony Group and Honda, with Fujitsu and Rakuten reportedly weighing whether to join. SoftBank engineers are working alongside researchers from Preferred Networks and AIST itself. It’s a familiar shape for a Japanese industrial push: rather than one company chasing a frontier model alone, the state has assembled a consortium of firms that already build the hardware this model needs to run on, from Honda’s robotics to Sony’s imaging sensors. Why robots, and why now Industry minister Ryosei Akazawa has been direct about the reasoning. The plan, he said, will “vigorously promote social implementation” across sectors, including restaurants, food manufacturing and medical care. Behind that language is a labour market running out of people: Japan’s ageing population, combined with tight migration policy, has left large parts of the economy short of workers with no easy fix in sight. Japan isn’t starting from nothing here. The country has spent years building robotics expertise in elder care, disaster response, manufacturing and even the Fukushima Daiichi cleanup. This project is an attempt to turn that experience into something exportable, not just a domestic patch. The timing also isn’t a coincidence. South Korea announced its own robotics push within a day of Japan’s confirmation, and both governments are framing physical AI as the next front in a competition that’s mostly been fought over chatbots and cloud contracts until now. What to watch next The real test isn’t the 2040 target, it’s the first stage-gate review. If Noetra hits its early milestones and releases a usable model this fiscal year, expect the investor list to grow well beyond the current four. If it doesn’t, the funding structure gives Tokyo every reason to walk away quietly rather than prop up a stalled national project. See also: From cloud to factory – humanoid robots coming to workplaces Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Japan’s answer to its worker shortage: An AI model for 10 million robots appeared first on AI News. View the full article

The Bank of England is reviewing whether existing rules can cover the use of agentic AI in finance, including payments, trading, cybersecurity, and operations. Deputy Governor Sarah Breeden said existing regulatory frameworks were not designed for AI agents that can act without direct human instruction. Speaking at the European Central Bank Forum on central banking in Portugal, she said relying on human oversight for every action by these systems is unlikely to be practical. Breeden said current frameworks were not built to contemplate autonomous agents in payments, trading, and operational functions. Agentic AI enters financial workflows Agentic AI refers to systems that can make decisions and carry out tasks independently. In finance, such systems are already being used in areas such as product recommendations, operational workflows, and trading-related tasks. Agentic systems differ from traditional automated trading tools because they can pursue objectives and make decisions with less direct human supervision. Breeden said these systems could act in similar ways if they are trained on similar data or designed around similar goals. Breeden said recent advances in AI models for identifying cyber vulnerabilities show a change in capability. She said agentic AI systems can chain together sequences of actions at scale and speed. A 2026 Cambridge Centre for Alternative Finance report found that 81% of surveyed financial services firms are adopting AI at some level. It also found that 52% of industry respondents are already actively adopting agentic AI. The report said most current use remains focused on internal functions, including process automation, data visualisation, software engineering, and knowledge management. Breeden said use in trading is still mostly concentrated in lower-risk operational tasks. BoE flags cyber resilience risks Breeden described cyber resilience as one of the Bank of England’s closest financial stability concerns around agentic AI. She said the technology has undergone a “step change” in cyber capability and that supervisors need to look at risks across the financial system rather than only at individual firms. She said AI tools can strengthen cyber defences when used by security teams. The immediate risk, she added, is that the same tools could increase the chance of attacks that harm financial stability if used by malicious actors. Breeden also noted that open-source models may trail the most advanced closed models by only four to eight months. She said this gives authorities only limited comfort, despite restrictions on the release of some advanced models. The IMF has also warned that AI-enabled cyber risk should be treated as a financial stability issue. It said attacks can scale quickly, spread across sectors that share digital infrastructure, and create wider disruption if several institutions are affected at once. Breeden said authorities should place greater weight on simultaneous disruption across several firms and stress-test the likely impact before such events occur. She said recovery planning may also need to account for mass disruption, rather than only isolated outages. The Bank of England is considering stronger recovery requirements for core systems. One option is to allow one bank to take over another bank’s basic functions during an outage or failure. Other options include arrangements that allow critical services to continue if a firm’s core systems are compromised. Breeden also raised the question of whether key firms should have separate failover systems or the ability to rebuild compromised core systems quickly. Tobias Adrian, financial counsellor and director of the International Monetary Fund’s capital markets department, also said AI poses serious risks to cyber resilience, according to Central Banking. The IMF has separately warned that shared software, cloud services, payment networks, and data networks can create correlated failures if widely used systems are targeted. Regulators consider market safeguards Breeden said regulators are also looking at guardrails, circuit breakers, and kill switches. These tools would be designed to limit or stop trading across markets if faulty AI models contribute to severe disruption. Breeden said autonomous systems could amplify volatility if they respond in similar ways to the same market signals, especially if their objectives drift from their original purpose or from public policy goals. The Bank of England has previously said existing rules were sufficient to manage AI-related risks. Breeden said recent developments have exposed gaps in current frameworks. Global regulators review AI safeguards The Financial Stability Board said earlier in June that AI agents pose a distinct challenge for human oversight and called for stronger safeguards. The FSB’s June consultation set out 12 proposed sound practices for responsible AI adoption by financial institutions. The practices cover organisation-wide governance, AI risk management across development and deployment, and AI-related cyber, ICT, and third-party risks. The FSB said the practices are not intended to create a binding international standard. It also said firms should define clear roles and responsibilities when using AI, especially when the technology is used in critical or material functions. Breeden said the Bank of England’s focus is on ensuring that financial firms remain resilient as autonomous systems are used in more areas. The review covers firm-level controls and market-wide safeguards. See also: HSBC expands AI banking partnership with Google Cloud Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Bank of England reviews AI rules for agentic AI in finance appeared first on AI News. View the full article

Anthropic has launched Claude Sonnet 5 and restored access to its Fable and Mythos frontier models following a federal export control review. The decision marks the conclusion of an eighteen-day operational pause triggered by a US government export control directive on June 12, which forced the temporary suspension of Anthropic’s highest-capability systems. Government officials enacted the restriction after researchers at Amazon documented a method to bypass the safety controls of Fable 5, causing the model to identify software vulnerabilities and supply exploitation code. Anthropic has since developed an updated automated classifier to patch the vulnerability, clearing the path for a full commercial rollout across its platform, cloud infrastructure, and partner networks. The temporary suspension of Fable 5 and Mythos 5 highlighted the regulatory pressures facing frontier intelligence systems. When the export control mandate took effect, the lack of real-time nationality verification systems required a total access blackout for all global users. Security evaluations conducted during the shutdown confirmed that the vulnerability identification behaviour was not unique to Fable 5. Older and less capable architectures from multiple providers, including Claude Opus 4.8, GPT-5.5, and Kimi K2.7, duplicated the exact results. To resolve the federal directive, engineers trained an automated safety classifier targeting the specific bypass mechanism reported by Amazon. This software layer functions with a wide safety margin, identifying and blocking ambiguous developer prompts that display a statistical probability of malicious intent. Internal validation data indicates the updated classifier prevents the reported exploitation technique in more than 99 percent of trials. When a developer issues a prompt that triggers this boundary, the platform automatically routes the workload to the older Opus 4.8 architecture to maintain continuity. The expanded safety margin introduces a distinct trade-off for engineering teams, as the automated system flags benign requests more frequently during routine application development and software debugging. Active deployments and agentic workflows While frontier models face strict state oversight, the immediate commercial focus targets the newly-deployed Claude Sonnet 5. Engineering teams are transitioning autonomous agents to this model to reduce operational expenditure while maintaining high execution capacity. Performance data validates that the system executes multi-step plans, operates terminal environments, and navigates web browsers without human intervention. Model performance and cost metrics: ModelSWE-bench ProTerminal-Bench 2.1Base input cost*Base output cost*Sonnet 563.2%80.4%$3.00$15.00Sonnet 4.658.1%67.0%$3.00$15.00Opus 4.869.2%82.7%$5.00$25.00 *Cost per million tokens. Sonnet 5 carries introductory rates of $2.00 input / $10.00 output through August 31, 2026. Real-world deployments demonstrate how organisations are deploying this architecture within live software development pipelines. At Rakuten, technology teams deployed the architecture against dozens of the company’s most challenging production code pull requests. The system processed each submission independently, executing tests and verifying the results before presenting the completed code to human engineers for final structural approval. Software automation firm Zapier integrated the system into its core product workflows to execute multi-part administrative tasks. In a documented deployment, engineers tasked the model with updating Salesforce account tiers and subsequently generating and transmitting launch announcements to enterprise contacts. Prior model architectures frequently stalled midway through these multi-stage operations, whereas the current system executed the entire sequence end-to-end without human remediation. Development tool provider Zed utilised the system to automate complex debugging procedures. During internal trials, engineering teams directed the model to investigate an active software bug. Working without explicit prompts or step-by-step instructions, the system independently generated a reproducing test script, applied the necessary code fix, and stashed the modifications to verify that the bug reappeared in the absence of the patch. The entire diagnostic and remediation sequence occurred within a single processing pass. Software engineering platform Factory implemented the architecture to manage sustained coding tasks within complex codebase environments. Technical teams reported that the system maintained logical grounding and execution consistency across corporate code repositories, outperforming previous generation software layers by completing tasks that previously timed out or failed to resolve. Quantitative safety audits and exploitation limits Data from the formal system card indicates that the system achieves these autonomous capabilities without a corresponding inflation of security risks. Automated behavioural audits designed to test for deceptive tendencies and cooperation with unauthorised requests show that the model exhibits a lower overall rate of non-compliant behaviour compared to its direct predecessor, Sonnet 4.6. The architecture does not possess advanced offensive cybersecurity capabilities. Anthropic engineers omitted specialised cybersecurity datasets from the training protocol, limiting the system to routine, defensive technical tasks. In public security assessments conducted in partnership with Mozilla, researchers tested the model’s capacity to build functional exploits for known vulnerabilities within the Firefox 147 browser core. The model failed to generate a single working exploit across all evaluation windows, registering a zero percent success rate. It did achieve a 13.2 percent partial success rate, which represented a minor increase over Sonnet 4.6, though engineers attribute this variation to general gains in logical reasoning rather than domain-specific offensive training. Out of caution, commercial versions ship with default real-time safety classifiers equivalent to those used in the premier Opus 4.8 framework. The regulatory friction surrounding Fable 5 prompted a formal partnership between Anthropic, Amazon, Microsoft, and Google to establish an objective industry framework for assessing model security breaches. Currently, providers lack a shared metric to classify the severity of system bypasses, creating regulatory uncertainty when researchers identify new prompting vulnerabilities. The proposed governance framework scores security breakdowns across four specific technical criteria: Capability gain measures how far the exploit advances user capabilities beyond standard, widely available software utilities. Breadth of capability gain quantifies the number of distinct offensive operations the same exploit unlocks. Ease of weaponisation tracks the volume of human engineering effort and specialized prompting required to extract a harmful output. Discoverability determines the accessibility of the exploit technique within public research circles. Developers and cybersecurity professionals will use this matrix to coordinate defensive responses. For high-severity breaches, such as exploits demonstrating an immediate capacity to disrupt financial accounting systems or electrical transmission grids, providers will deploy automated mitigations instantly. This initiative operates alongside a newly established HackerOne vulnerability research program and a dedicated corporate monitoring team providing 24-hour oversight of threat intelligence channels. Deployment strategies will need to adapt to this closer relationship between model builders and state regulatory bodies. Anthropic has formalised agreements under recent executive mandates to grant federal researchers early access to frontier architectures prior to public commercial release. These joint evaluation windows allow external security analysts to audit model capabilities alongside internal engineering teams, ensuring regulatory alignment before code enters production environments. See also: HP accelerates enterprise workflows with OpenAI Frontier Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Anthropic deploys Claude Sonnet 5, Fable and Mythos restored appeared first on AI News. View the full article

HP has scaled its OpenAI Frontier integration across global operations to optimise enterprise workflows and accelerate output. The hardware manufacturer initiated testing of the platform in February 2026. Early pilot programs yielded verified operational gains in software engineering and cybersecurity remediation. Expanding these initial trials into an enterprise-wide operating model requires connecting access protocols, contextual data, and evaluation metrics. Frontier supplies this connective tissue. Engineering capacity and deployment metrics Implementation metrics indicate high usage among technical staff. One HP engineer processed 122 pull requests spanning 43 distinct projects within a matter of weeks using OpenAI models. Managing pull requests across dozens of concurrent projects typically induces severe context switching penalties for human operators. Automated models process repository syntax and validate code logic across multiple environments simultaneously. This capability directly reduces wait states within the software development life cycle. The corporate security division applied these identical models to resolve several software bugs within a single day. Internal estimates indicated this remediation workload would typically consume an entire month. Enterprise development teams lose countless hours as code transitions through testing protocols, peer reviews, security audits, and sprint planning schedules. OpenAI tools compress these isolated stages into a collaborative and accelerated sequence. Technical execution speed increases when diagnostic tools accurately pinpoint flaws during initial commits. “It has been an amazing tool, and I am using it daily,” an HP engineer stated. The deployment architecture segments AI models based on task requirements. HP directs ChatGPT instances to execute broad knowledge initiatives. These implementations manage active enterprise research, data analysis routines, concept ideation, and automated workflow triggers. Codex instances handle specialised development operations. Engineers instruct Codex to map application planning phases, construct user interface scaffolding, and manage parallel software-delivery tasks. Separating workloads across designated models prevents processing errors and ensures accurate output generation. Partner channel integration External partner networks are the bulk of HP’s operational flow. More than 80 percent of the company’s business travels through its channel ecosystem. Over 100,000 partners globally access the HP Partner Portal. Applying AI to this massive external network requires strict data routing. Enterprise software ecosystems fail when partner portals experience lag or present inaccurate administrative data. The Frontier platform facilitates a cohesive self-service architecture covering store interfaces, partner communications, and voice channels. AI agents supply constant guidance regarding program navigation and business information. These agents process partner queries and deliver direct operational management support. This deployment decreases manual processing loads and accelerates information-to-action cycles. Customers and partners execute routine workflows and attain resolution faster through these automated systems. Administrative queries regarding stock limits or warranty routing resolve without human intervention. Device telemetry and fleet management Hardware administration relies on the HP Workforce Experience Platform (WXP). CIOs use this central dashboard to oversee entire device fleets. Processing device health signals across global corporate networks generates massive data payloads. Human technicians cannot manually correlate every error log across tens of thousands of deployed machines. HP integrates Frontier to analyse device telemetry, operational objects, schemas, and runbooks. AI agents process fleet health signals to investigate application hangs, Wi-Fi connectivity errors, and system crashes. This diagnostic speed promotes accurate remediation protocols across distributed corporate environments. The platform provides a single pane of glass for device management. Automated investigation of operational objects ensures hardware failures register immediately and map to established recovery procedures. IT teams can initiate repairs based on analysed telemetry rather than basic user complaints. Enterprises require agents that understand trusted context boundaries. Frontier provides the necessary connectivity to govern APIs and evaluate system outputs. Shadow IT environments develop when departments deploy unmonitored AI instances. Frontier centralises these deployments. Security operations serve as both an operational proof point and an active governance layer. HP security personnel deploy ChatGPT to proactively neutralise vulnerabilities. Directional estimates project roughly 82 hours per week of security-team capacity freed by this automation. Retaining cybersecurity professionals requires eliminating monotonous log review processes. Frontier maintains oversight by managing permissions, evaluation parameters, and deployment controls. Human capital executes higher-level analysis while automated tasks remain fully-reviewable. HP is not only optimising current operational capacity but establishing a robust framework for future technological integration—ensuring that as enterprise demands evolve, the underlying infrastructure remains secure and agile. See also: Wimbledon adds IBM AI tools for live match coverage Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post HP accelerates enterprise workflows with OpenAI Frontier appeared first on AI News. View the full article

The All England Lawn Tennis Club is adding new AI-powered features to Wimbledon’s digital platforms through its ongoing work with IBM. The updates will be available through the Wimbledon app and wimbledon.com as first-round matches begin on Monday. They include an upgraded Match Chat assistant and a new feature called Key Moments. The features will also be available through IBM Slamtracker, which is accessible on the Wimbledon app and wimbledon.com. CIO.inc reported that about 730 million people engaged with Wimbledon last year, generating 18 billion impressions across digital channels, according to comments from AELTC during an IBM media panel. IBM said its work with Wimbledon forms part of a five-year digital transformation project focused on modernising the digital platform, bringing critical services and data back in house, and reducing technical debt and cost of ownership. IBM said the work is intended to enable an AI operating model and increase the productivity of the Wimbledon team. Match coverage gets AI tools Match Chat allows fans to ask questions about a match in natural language. A user can ask what has happened so far in a match and receive a conversational response based on live data, analysis, and historical performance information. Some Match Chat responses will include photos and video where relevant. IBM said the feature lets users ask questions without searching through statistics or moving between multiple screens during a match. IBM said Match Chat is built on watsonx Orchestrate and uses AI agents and models trained on Wimbledon’s editorial style and tennis terminology. The assistant uses live match data, analysis, and historical performance information, with some responses including relevant photos and video. A 2025 technical paper on Match Chat reported that earlier deployments at Wimbledon and the US Open served about 1 million users, with an average response time of 6.25 seconds. The 2026 version adds expanded data sources and selected responses with photos and video, according to IBM. The new Key Moments feature identifies points and passages of play that influence the direction of a match. It explains pivotal plays and changes in momentum using AI-generated analysis. Key Moments builds on Wimbledon’s existing Likelihood to Win feature. That tool calculates each player’s probability of winning throughout a match by using current and historical statistics, expert input, and match momentum. IBM said Live Likelihood to Win combines scoring data, match statistics, and expert analysis into continuously updated win probabilities for every singles match. Key Moments builds on that system by identifying points and passages of play that affect the direction of a match. Key Moments will be available for each gentlemen’s and ladies’ singles match. IBM said the feature is intended to explain which plays influence the direction of a match and why, rather than only showing selected points. IBM said Key Moments can identify examples such as long rallies or double faults when they affect match momentum or win probability. Inside Wimbledon’s platform rebuild The digital updates follow a broader modernisation of Wimbledon’s app and website. As part of that work, Wimbledon’s content archive was moved to a new architecture covering more than 15,000 digital assets. The archive includes articles, videos, photographs, and metadata links used across Wimbledon’s digital services. IBM said the platform uses watsonx Orchestrate for AI agents, IBM Bob for development work, and watsonx.data to manage data across hybrid cloud infrastructure. The redesigned app and website were also built around research into how fans, players, broadcasters, media, and other groups use Wimbledon’s digital services. IBM Bob was used to build a knowledge graph mapping relationships across Wimbledon’s archive and to support AI-driven workflows for the new platform. Computer Weekly reported that the tool examined around 15,000 online assets and learned how articles, photos, videos, and metadata were connected. IBM said the archive mapping work, which it said could traditionally require four to five IBM specialists working for months, was completed by one engineer within four weeks. The targeted 15,000 assets were extracted in 47 minutes, although IBM noted that response times can vary depending on technical conditions. Computer Weekly also reported that IBM Bob helped complete what was described as 10 years of development work in nine months during the rebuild of Wimbledon’s app, website, and content systems before The Championships 2026. Controls, engagement, and history CIO.inc reported that IBM and Wimbledon are also using governance controls for the AI features. The report said the design includes human-led processes, explainability, confidence scoring, and checks intended to reduce inaccurate outputs during live use. Insider Sport separately reported that IBM Consulting pointed to a watsonx governance layer across the features, with transparency measures and confidence scores attached to the underlying data. AELTC reported a 16% year-on-year increase in engagement across all platforms in 2025 and a 39% rise in myWIMBLEDON registrations over the past year. IBM and the All England Club have worked together for more than 35 years. Their partnership has covered Wimbledon’s website, mobile app, and AI-powered digital services. IBM said the partnership has included the launch of the Wimbledon website in 1995, the mobile app in 2009, and the first integration of enhanced AI-powered solutions in 2017. (Photo by Gonzalo Facello) See also: The math behind the OpenAI Jalapeño chip Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Wimbledon adds IBM AI tools for live match coverage appeared first on AI News. View the full article

Natural language processing is reshaping professional communication on online platforms, enabling more relevant and personalised networking interactions. As AI-driven systems increasingly comprehend and generate human language, these technological advances affect how users pursue and maintain professional connections, presenting both opportunities and challenges in authentic relationship-building. Professional networking now relies on a growing ecosystem of AI-powered tools that affect how people initiate and manage online connections. New language technologies are streamlining introductions, fostering more contextual conversations, and helping users navigate crowded digital spaces. LinkedIn lead generation services are evolving as modern natural language processing systems interpret user intent and enhance the relevance of initial outreach. By automating repetitive communication tasks while keeping human intent central, these innovations adjust strategies and reshape expectations for network development. Core features of NLP transforming online networking Contemporary natural language processing introduces essential capabilities that make connecting online more efficient and engaging. AI systems now offer message intent detection, enabling automation tools to filter out unsolicited contact and present users with relevant opportunities. These models detect details in language that reveal professional interest, urgency, or polite disengagement, helping reduce unwanted messages and improving conversation quality. Flexible tone and style adaptation is another notable advance, allowing digital correspondence to align with varied industries, cultures, or organisational hierarchies. Summarisation tools can condense lengthy profiles or conversation histories, enabling users to quickly evaluate potential connections or remember the context of ongoing discussions. Additionally, multilingual support in natural language processing further facilitates networking across language barriers while preserving nuanced meaning. Semantic matching elevates connection relevance and fit Earlier professional networking platforms often matched users by keywords, linking job seekers, recruiters, and collaborators based on basic criteria. In contrast, current networks leverage semantic matching and embeddings to assess the intent and meaning behind interactions. This shift allows for more targeted connection recommendations and helps minimise ineffective mass outreach by emphasising shared context, experience, or objectives. Natural language processing tools can analyse conversational content for signs of overlapping interests and relevant experience, moving beyond skills or job titles alone. Connections may be suggested based on shared projects, mutual challenges, or similar goals. This approach can lower the rate of irrelevant messages and foster more meaningful engagement, aligning with user priorities. Balancing personalisation, authenticity, and digital safety Advanced natural language processing increases personalisation in networking encounters. Messages customised to context and conversation history can enhance relevance, meet professional norms, and boost user satisfaction. However, when AI-generated messages mimic personal details too closely or contain false information, they may seem intrusive or undermine trust. One ongoing concern is inaccuracies—called hallucinations—where AI systems generate convincing but false information, which can be problematic in professional contexts where reliability matters. Language bias within natural language processing models can also introduce stereotypes or unsuitable tones, particularly when algorithms suggest connections or draft outreach messages. As a result, there is a growing emphasis on transparency and verifiable claims to maintain honest communication, prompting users to be vigilant for inconsistencies in generated content. Shaping data policy and anticipating future networking trends Advancements in digital networking increasingly intersect with privacy standards, consent regulations, and policy requirements. Managing data collection and processing for natural language processing applications requires strong governance to align with user expectations and comply with laws. Tools are being designed to balance improvements in networking capabilities with the importance of privacy and informed consent. Professionals are also watching for innovations such as specialised smaller language models capable of operating on device or with robust privacy protections. Developments like privacy-preserving inference, enhanced fact-checking, and better evaluation metrics are being integrated to align natural language processing with the needs of professional networking. As such capabilities expand, emphasis is expected to remain on building relationships while sustaining transparency and trust among users. The post Advances in Natural Language Processing Are Changing Professional Networking appeared first on AI News. View the full article

Modern DevSecOps needs security checks that run before release day. Teams now write code, build services and deploy updates at a pace that manual review cannot match. That’s why they use automated testing, as it helps catch routine flaws before they reach production. The pressure has grown. Verizon’s 2025 Data Breach Investigations Report found that vulnerability exploitation caused 20 percent of breaches as an initial access route, up 34 percent from the prior report. It also found that credential abuse caused 22 percent, which shows why code flaws and access flaws need attention together. Automated testing has become more valuable as software teams release changes faster. Services like XBOW support that work by mapping application surfaces, testing likely attack routes and validating whether a finding can lead to real access. For security professionals, the benefit lies in better proof, fewer vague tickets and faster handoffs to engineering teams. Start with code testing Static application security testing checks source code before the software runs. It can find weak input handling, unsafe functions and risky patterns in pull requests. Developers value this because the test happens near the line that caused the issue. Nobody enjoys reopening a ticket three weeks after the code has travelled through six approvals. Static testing works best when teams tune rules. A scanner that flags every minor issue will lose trust. A good setup focuses on high-risk patterns, clear fixes and ownership. OWASP’s DevSecOps guidance places security testing inside the pipeline so teams can find issues during development instead of waiting for a later review. Test the running application Dynamic application security testing checks a live application from the outside. It sends requests to a running service and looks for unsafe responses. This helps teams find flaws that code review may miss, such as broken access checks or unsafe redirects. Dynamic testing needs care because it touches real systems. Teams should test staging environments where possible, set safe limits and record what the tool did. The value comes from proof. A finding that shows the tested request, the response and the affected route gives developers a concrete starting point. Platforms like Xbow fit this part of the toolset when teams need automated penetration testing for web applications. The platform describes controlled, non-destructive validation before surfacing findings, which supports a stronger link between test output and real exploitability. Check dependencies before they check you Software composition analysis reviews third-party libraries and open-source packages. That matters because most modern applications depend on code that no internal team wrote. A package can save time, but it can also bring a known flaw into a build. CISA’s Known Exploited Vulnerabilities catalog gives teams a practical source for prioritising flaws that attackers have used in the wild. Security teams should use that kind of evidence when they decide which dependency updates need urgent work. Dependency testing should run in pull requests and scheduled checks. A project may pass today, then become exposed next month after a new advisory. Automated checks help teams catch that change without asking someone to reread every package list by hand. Protect secrets and build settings Secret scanning checks code and configuration for passwords, tokens and keys. This has become a basic need because one exposed token can give an attacker access without a software bug. A 2025 report from TechRadar described research that found more than 17,000 exposed secrets across public repositories and indexed web data. Infrastructure-as-code testing checks cloud templates and deployment files. In plain terms, it looks at the instructions that build servers and services. This can catch open storage, weak identity rules and risky network settings before deployment. The best tests show both the risky line and the safer option. Use AI with limits Advancements in AI have led automated testing has started to move from pattern matching toward reasoning. AI can help tools explore more paths, draft clearer remediation notes and test combinations that older scanners may miss. It can also create confidence that the evidence has earned. That promise needs discipline. The Guardian reported in May 2026 that Google had warned about AI-powered hacking reaching industrial strength, with criminal and state-linked actors using advanced models to improve malware and exploit work. Defensive teams therefore need automation that can keep pace, but they still need humans to approve scope and judge impact. Modern platforms, including Xbow, use AI to simulate attacker behaviour across web targets and then validate findings before reporting them. That supports DevSecOps teams that need faster tests without turning every alert into a meeting. The right outcome is fewer unclear findings rather than more alerts. Prioritise attack paths Many teams still rank issues by severity score alone. That can mislead. A medium issue that links to exposed credentials may matter more than a severe issue blocked by access controls. Attack path analysis looks at how flaws connect. This approach helps business leaders understand risk. They need to know whether an attacker can reach customer data, change production code or take over an account. A good automated tool should make that path visible and show the control that breaks it. IBM’s 2025 Cost of a Data Breach Report put the global average breach cost at $4.44 million. That number gives leaders a reason to fund testing, but the daily work still comes down to fixing reachable risks before attackers use them. The post Best Automated Security Testing Tools for Modern DevSecOps appeared first on AI News. View the full article

xFusion presented scalable enterprise AI computing models at ISC 2026, transitioning hardware from edge devices to data centres. Enterprise technology buyers attending the Hamburg exhibition sought practical production frameworks. Hardware selection processes regularly fail to account for physical operating limits. Relying on public APIs exposes proprietary commercial data. xFusion engineers responded with a four-tier hardware portfolio. The deployment structure scales processing capacity sequentially through individual workstations, workgroup clusters, corporate office appliances, and facility-level supernodes. Personal edge processing devices Engineers and specialised staff require dedicated local resources to execute complex 3D rendering and architectural simulation tasks. Individual professional users process vast datasets locally before committing workloads to centralised computing clusters. xFusion supplies the FusionXtation X3 8000 Gen2 edge computing node as the foundation layer. Corporations deploy these workstations to staff requiring local execution of 70-billion to 200-billion parameter models. Hardware configurations pair Intel Core Ultra processors with dual professional-grade graphics processing units. Memory configurations include error-correcting DDR5 RAM up to 256 gigabytes and internal storage reaching eight terabytes. Production environments report a 70 percent faster 8K rendering output and up to a 50 percent boost in general AI processing performance compared to previous hardware iterations. IT administrators maintain remote access through integrated Baseboard Management Controllers. Four 40-gigabit-per-second Thunderbolt ports handle external data transfers. Workgroup data containment appliances Uncontrolled data flows present compliance risks to regulated institutions. External application marketplaces expose corporate networks to malicious code. Development teams must construct custom software securely to protect corporate intellectual property. xFusion designed the FusionXpark appliance to service this team-level compliance requirement. The platform equips engineering units to maintain regulatory standards during initial application design. Medical imaging teams and financial modellers process highly-sensitive commercial data entirely isolated from external APIs. Teams combine two independent FusionXpark units to process 405-billion parameter models locally on native CUDA environments. The system runs NVIDIA DGX OS directly from the factory. Developers gain immediate access to required toolchains. Network administrators route overflow processing demands securely into DGX Cloud through native integrations. Corporate token processing utilities High-volume corporate functions consume processing capacity at unsustainable rates. Redundant context transmission inflates operational budgets. Operations departments require predictable infrastructure to execute automated customer service routines and process complex financial approvals. xFusion manufactures the TokenBox to act as a centralised corporate processing appliance. The unit supplies token generation capacity across entire companies from a single on-premises installation. Individual TokenBox nodes hold enough capacity to run models with 1.6 trillion parameters. The deployment model avoids the capital expense of constructing dedicated server rooms. Internal data-centre-grade liquid cooling mechanisms restrict noise levels to 35 decibels during active computations—enabling facility managers to place these units directly into normal office environments. Pre-installed software and skills bypass lengthy configuration periods to accelerate return on investment. Data centre computing engines Scaling computing capabilities across multinational corporate networks requires advanced thermal infrastructure. The final deployment tier focuses entirely on liquid-cooled racks and supernodes. High-density infrastructure packages manage 240 kilowatts per cabinet. Hardware designs translate raw electrical input directly into floating-point operations per second. Self-developed low-loss core components reduce single-module operational expenditures by 15 percent. Centralised corporate clusters rely on dense rack-mounted hardware integrated within this supernode architecture. The FusionServer G6550 V8 inference server operates within this tier, housing up to ten dual-width graphics processing units. The FusionPoD liquid cooling computing platform manages the thermal output of these racks, achieving a partial Power Usage Effectiveness of 1.06. Liquid cooling loops extract heat using graphene pads and diamond cold plates that reach conductivity ratings of 1200 watts per metre-kelvin. Processing hardware stalls without constant data ingestion. The FusionOne DFS storage solution provides the underlying data architecture for these compute clusters. A three-node cluster operating 72 NVMe drives achieves sequential read bandwidths of 200 gigabytes per second. The arrays scale to manage exabytes of total capacity while achieving 94.1 percent storage utilisation through erasure coding. Executives assessing infrastructure deployment frameworks can access technical product specifications directly at [Hidden Content]. The post xFusion scales enterprise AI from edge workstations to liquid-cooled data centres appeared first on AI News. View the full article

New partnership brings on-device deepfake detection to video calls on desktop SAN FRANCISCO, June 29, 2026 — Scam.ai today announced a partnership with Qualcomm and the launch of Halo, an on-device deepfake detection model for live video calls. Both announcements were made at Computex 2026 in Taipei, where Scam.ai was featured at Qualcomm’s booth as part of the event’s Agentic AI track, showcasing the latest in on-device deepfake detection technology. “Deepfakes has rapidly risen as one of the major concerns for enterprises in multiple fronts, all traditional security measures are oblivious once a human is breached,” said Dennis Ng, co-founder of Scam.ai. “By checking the video securely and privately on-device, we’re able to curb attacks from the source ” The Qualcomm partnership gives Scam.ai access to device ecosystem resources and optimization support, enabling Halo to run locally on personal computers without relying on cloud infrastructure. Halo operates in the background of any video conferencing session and flags synthetic or AI-generated video in real time. The product is optimized for Qualcomm-powered devices and runs on desktop computers. HR and recruiting teams conducting video interviews, and high-value executives including CEOs, CFOs, and venture capitalists who take frequent, high-stakes calls. Identity fraud in video interviews has become a documented enforcement risk, with only 31% of HR leaders reporting they feel equipped to detect it, and deepfake fraud attempts increasing over 2,000% in the past three years. Key capabilities: Real-time detection: Identifies synthetic video during a live call, not after On-device processing: No video footage leaves the user’s computer Passive operation: Runs in the background with no change to existing call workflows Halo is available June 2026. Enterprise integration details and additional platform partnerships will be announced in the coming months. About Scam.ai Scam.ai is the real-time deepfake detection platform trusted by enterprises to secure their most critical interactions. Powered by a proprietary on-device AI engine, Scam.ai helps organizations detect synthetic and AI-generated video before fraud occurs. Across hiring, executive communications, and high-stakes financial calls. Headquartered in San Francisco and backed by world-class investors, Scam.ai is on a mission to make every human interaction trustworthy. To learn more, visit [Hidden Content]. MEDIA CONTACT Neo Tiangratanakul *****@*****.tld The post Scam.ai Announces Qualcomm Partnership, Launches Halo Deepfake Detection Model at Computex 2026 appeared first on AI News. View the full article

SAP aligns fragmented commerce data structures to enable operational AI personalisation at the execution layer. Enterprise leadership routinely establishes objectives to anticipate customer requirements and deliver relevant interactions across digital touchpoints. However, the actual infrastructure running inside these enterprises fails to support systematic execution at the required volume. Recommendation engines display generic product listings because the underlying behavioural data remains isolated. Marketing departments dispatch email communications based on rigid calendar schedules rather than adapting to individual user habits. Corporate loyalty programs issue rewards based entirely on financial transactions while ignoring broader relationship metrics. The technical ambition exists, yet the foundational architecture remains incomplete. Clean data resides in disconnected repositories. AI capabilities sit dormant within the technology stack. Organisations lack the operational discipline required to execute continuous experimentation. SAP engineered the ‘Advanced Success Plan’ for SAP Customer Experience solutions to resolve these deployment failures. Three layers of advanced AI personalisation System architects cannot activate advanced personalisation through standard configuration switches. Enterprise implementations require systematic construction across three connected operational layers encompassing data, decisioning, and delivery. Data serves as the required baseline architecture. Enterprise systems must aggregate unified, real-time customer profiles while maintaining strict consent awareness. These profiles consolidate information from completed commerce transactions, historical engagement records, active browsing behaviour, customer service tickets, and ongoing loyalty activity. AI models require these complete behavioural data points to function; without this aggregated data, the algorithms operate on defective inputs. The decisioning layer processes these behavioural data points into executable directives. AI algorithms evaluate the incoming data streams to determine the optimal next product to display, select the exact promotional offer to present, and calculate the precise moment to initiate contact. This layer demands rigorous governance frameworks. System administrators must define operational parameters dictating when the automated algorithm controls the output and when human operators override the machine logic. The delivery layer executes the personalised experience and presents it to the customer. The system transmits these tailored interactions through the digital storefront, directly into email inboxes, via mobile push notifications, and across loyalty program interfaces. Enterprise architecture requires precise orchestration across these channels to ensure the outgoing communication matches the customer’s live context. The Advanced Success Plan targets these three layers simultaneously, deploying expert technical guidance and governance structures to transition organisations away from disconnected point solutions toward an integrated operating model. SAP Commerce Cloud storefront execution mechanics SAP Commerce Cloud operates as the storefront execution engine for large-scale personalisation. The software features an AI-assisted product recommendation system that displays relevant inventory to individual visitors at precise moments during their shopping sequence. The engine surfaces trending merchandise, related catalogue items, and complimentary accessories designed to drive cross-selling and upselling metrics. The system bypasses static manual merchandising configurations to evaluate real-time behavioural inputs. This automated evaluation improves conversion performance and increases product discovery at a volume that human merchandising teams cannot manually replicate. Administrators running SAP Commerce Cloud often fail to activate these advanced features due to predictable technical barriers. Deficient data quality degrades the accuracy of the recommendation models. Integration complexities sever the data connections between the storefront application and the upstream customer profile databases. Marketing departments lack the internal testing frameworks necessary to tune and optimise the algorithms. The Advanced Success Plan deploys targeted technical interventions to clear these blockages. Technical teams execute data readiness assessments to measure baseline information quality and map the integration pathways required to transmit clean behavioural data into the personalisation engine. Adoption accelerators install structured testing workflows, allowing marketing operators to define hypotheses, execute A/B tests, and write successful modifications into permanent platform configurations. The result is that the digital storefront evolves into an adaptive system that learns from incoming data rather than operating on static initial settings. Automating customer lifecycles via SAP Engagement Cloud SAP Engagement Cloud, powered by the SAP Emarsys platform, pushes this personalisation framework past the digital storefront and across the complete customer lifecycle. The system ingests transactional data from SAP Commerce Cloud and merges it with historical engagement records to generate cross-channel communications targeting individual users rather than broad audience segments. The AI-assisted send time optimisation feature executes this individualised approach. The algorithm abandons fixed transmission schedules to analyse the unique behavioural patterns of every single contact. The system ignores standard time zone, language, and regional constraints to dispatch messages at the exact second the individual user demonstrates the highest statistical probability of engagement. This process automates personalised communication into a scalable operational workflow. Marketing departments pair this optimisation tool with the SAP Emarsys AI-assisted campaign translator and omnichannel orchestration systems to abandon static campaign creation. Teams orchestrate dynamic automated journeys where the software continuously evaluates which user actions should activate specific communications. The system modifies these interactions based entirely on response metrics. The native technical integration connecting SAP Commerce Cloud and SAP Engagement Cloud accelerates the deployment timeline. Merging commerce activity with external engagement data increases overall conversion rates, elevates purchase frequency, and expands the average order value. Independent, disconnected systems cannot achieve these financial metrics. The Advanced Success Plan secures this joint platform value by coordinating the integration architecture, establishing data governance protocols, and tracking adoption milestones across both environments. Implementing outcome-based governance models Teams routinely misclassify personalisation initiatives as single-phase software implementations. The SAP framework restructures these deployments into continuous improvement operations. SAP’s plan enforces outcome-based governance by establishing target KPIs. Stakeholders track conversion rate lift, track repeat purchase volume, monitor engagement open rates, and calculate average order values. Project managers build dedicated work streams designed to advance those metrics. Implementation specialists follow prescriptive adoption patterns organised into structured playbooks. These manuals provide the technical steps required to activate AI-assisted recommendations, configure send time optimisation logic, and deploy next-best action algorithms through quantified gates. The program delivers continuous role-based enablement and coaching directly to data engineers, product owners, and campaign managers. This targeted training closes internal skills gaps that typically cause personalisation operations to stall or regress. Proactive telemetry systems keep tabs on the live deployment. Automated adoption checks scan the platform to identify underperforming configurations. AI-guided best practice alerts inform system administrators about necessary tuning adjustments before poor configuration impacts enterprise revenue. The financial justification for these system upgrades relies entirely on verifiable operational data. SAP Commerce Cloud administrators track the value of operationalised hyper-personalisation through direct storefront metrics. Upgraded systems report higher transaction conversions generated by AI-surfaced recommendations, increased average order values secured through automated cross-selling, and improved product discovery rates that lower site abandonment. SAP Engagement Cloud operators measure system value through communication quality metrics. Upgraded systems record higher open and click-through rates driven by individual user relevance. Automated delivery timing improves overall campaign return on investment. Loyalty programs generate deeper interaction metrics based on relationship strength rather than simple transaction volume. The integration of unified data and automated decisioning restructures hyper-personalisation from a static proof-of-concept into an automated financial growth mechanism that measurably improves over time. See also: Omio scales travel product development using OpenAI models Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post SAP aligns commerce data for AI personalisation appeared first on AI News. View the full article

OpenAI’s financial trajectory hinges heavily on infrastructure costs, a reality that drove the development of the new custom OpenAI Jalapeño chip. Developed in collaboration with Broadcom, the application-specific integrated circuit (ASIC) represents a direct attempt to mitigate the heavy capital expenditure associated with third-party hardware. While Nvidia currently commands an estimated 75% profit margin on its high-end processors, OpenAI operates on tighter margins, keeping roughly 33 cents of profit on each dollar generated after accounting for its massive operational expenses. The financial burden of running large language models at scale is severe. Last year, keeping ChatGPT servers responsive had cost OpenAI a staggering US$8.4 billion. With the platform now attracting 900 million weekly users, that operational cost is projected to reach approximately US$14 billion this year. Over the next eight years, OpenAI has committed roughly US$1.4 trillion to computing power, a massive bet for a company currently generating US$25 billion in annual revenue. Designing Hardware for LLM Inference The OpenAI Jalapeño chip, dubbed as the company’s first “Intelligence Processor”, is built specifically for large language model (LLM) inference rather than general-purpose AI workloads. OpenAI provided the core architectural design based on its specific model roadmaps and serving systems, while Broadcom managed the silicon engineering and high-performance networking integration. TSMC handles the physical manufacturing in Taiwan, and Celestica is tasked with building the board and rack systems. According to OpenAI, early lab samples are already running frontier workloads, including an unreleased GPT-5.3-Codex-Spark model, at target production frequency and power. Richard Ho, head of OpenAI’s hardware program, noted that the architecture minimizes data movement to push realized utilization closer to its theoretical peak performance. Unlike general-purpose accelerators adapted from legacy AI workloads, this architecture specifically balances compute, memory, and networking resources to solve the data-movement bottlenecks native to interactive LLM serving. To achieve this at scale, the platform integrates Broadcom’s Tomahawk networking silicon directly into the design, allowing the custom processors to communicate across massive, clustered data center environments. The vertical integration flywheel By moving into custom silicon, OpenAI shifts from being a mere software layer to a vertically integrated infrastructure company. This full-stack strategy spans the entire pipeline: chip architecture, software kernels, memory systems, network scheduling, and the final application layer. Much like Apple’s tight coupling of proprietary hardware and iOS, OpenAI can now optimize its infrastructure around its exact internal model roadmaps. This integration feeds a continuous operational flywheel. Enhanced infrastructure efficiency lowers the cost of both training and serving models. More affordable serving leads to better, more responsive products, which drives user volume and revenue to be reinvested back into the next generation of custom infrastructure. Overcoming the late-mover advantage By introducing its own silicon, OpenAI enters a landscape where its primary competitors have spent nearly a decade developing proprietary hardware. Google began deploying its Tensor Processing Units (TPUs) in 2015 and now controls roughly a quarter of global AI computing capacity outside of Nvidia’s supply chain. Amazon has shipped over one million of its custom chips, while Meta and Microsoft continue to scale their own infrastructure. “Jalapeño is part of our long-term full-stack infrastructure strategy to make compute more abundant,” said Greg Brockman, president and co-founder of OpenAI. “By designing more of the stack ourselves, we can serve more intelligence with greater efficiency.” To close this timeline gap, OpenAI accelerated the development phase. The OpenAI Jalapeño chip transitioned from a blank-slate design to manufacturing tape-out—the final step before physical production—in just nine months. The engineering teams achieved this timeline by utilizing OpenAI’s own language models to automate and optimize portions of the hardware design process. This creates a unique feedback loop where the models served to users are actively being leveraged to build the physical infrastructure that will run future iterations. Initial deployment of the hardware into data centres is scheduled to begin by the end of 2026. Broadcom CEO Hock Tan confirmed that the rollout will scale alongside infrastructure partners, including Microsoft, to prepare for gigawatt-scale data centre integration. (Photo by OpenAI) See also: Omio scales travel product development using OpenAI models Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post The math behind the OpenAI Jalapeño chip appeared first on AI News. View the full article

Samsung Electronics is expanding employee access to ChatGPT Enterprise and Codex, giving staff wider use of AI tools for technical and non-technical work. According to OpenAI, the deployment covers all Samsung Electronics employees in Korea and all Device eXperience employees worldwide. The DX division includes smartphones, consumer electronics, and home appliances. Samsung plans to use the tools in software development, marketing, product development, manufacturing, and other business functions. The tools will support tasks such as information search, document drafting, idea development, data interpretation, and code-related work. Samsung revisits employee AI use The rollout comes three years after Samsung restricted employee use of generative AI tools over data-security concerns. In 2023, the company limited the use of ChatGPT and similar tools after concerns that sensitive internal information had been uploaded to an external AI platform. The new deployment gives employees access to ChatGPT Enterprise, which includes controls for data protection, user access, and security management. OpenAI said the enterprise version allows organisations to manage users, apply access controls, and use AI tools within internal security requirements. Samsung’s earlier restrictions applied to employee use of ChatGPT and similar generative AI tools. The new rollout gives employees access through an enterprise product with data protection and access controls. Samsung has not limited the deployment to a single business unit or technical group. OpenAI said the tools will be used across a broad range of functions, including technical and non-technical teams. OpenAI said ChatGPT can support knowledge-based tasks such as searching for information, analysing material, drafting documents, developing ideas, and interpreting data. Codex for technical and non-technical work Codex will be used for software-related tasks such as writing, reviewing, and debugging code. OpenAI said the tool is also being used for internal tools, websites, software prototypes, and automated workflows. OpenAI said Codex can also support non-technical teams in day-to-day work, including by helping employees create internal tools and automated workflows. OpenAI said Codex now has more than five million weekly users across technical and non-technical workflows. In Korea, weekly active users of Codex have grown nearly 800% since February 1, 2026, according to the company. Harrison Kim, general manager of OpenAI Korea, said the agreement is one of OpenAI’s largest enterprise deployments. He said Samsung is using AI across teams and functions rather than limiting it to specific departments. Samsung and OpenAI’s infrastructure links In October 2025, Samsung said it would work with OpenAI as a strategic memory partner for the Stargate AI infrastructure initiative, with OpenAI’s memory demand projected to reach up to 900,000 DRAM wafers per month. Samsung SDS also entered a potential partnership with OpenAI to jointly develop AI data centres and provide enterprise AI services. Samsung said the agreement would allow Samsung SDS to provide consulting, deployment, and management services for businesses integrating OpenAI models into internal systems. Samsung SDS also signed a reseller partnership to offer OpenAI services in Korea. Under that arrangement, Samsung SDS said it would support Korean companies adopting ChatGPT Enterprise and other OpenAI services. Reuters reported that Samsung Electronics and SK Hynix had signed letters of intent to supply memory chips for OpenAI’s Stargate project. The report said the two South Korean chipmakers together account for about 70% of the global DRAM market and nearly 80% of the high-bandwidth memory market. High-bandwidth memory supports fast data movement between memory and processors in AI systems. Reuters reported that OpenAI’s chip demand for Stargate may reach 900,000 wafers per month, citing South Korea’s presidential office. Samsung said its semiconductor businesses would support OpenAI’s demand with advanced memory solutions. The company also said its affiliates were exploring broader work with OpenAI in areas including data centres, enterprise services, and AI infrastructure. AI adoption and productivity Deloitte’s 2026 State of AI in the Enterprise report found that 66% of organisations reported productivity or efficiency gains from enterprise AI adoption. The same report found that 53% reported improved insights and decision-making. A Bpifrance survey reported by Reuters found that 77% of 534 French mid-sized company heads said their firms used generative AI, but only 17% of those using it reported time savings. Samsung has identified use cases across document work, information analysis, coding, product development, marketing, and manufacturing. The deployment gives employees access to ChatGPT Enterprise and Codex for those tasks under a company-wide agreement. OpenAI’s Korea partnerships OpenAI has also announced other partnerships in Korea. Seoul National University recently began providing ChatGPT Edu to 47,000 students, faculty, and staff. OpenAI has also worked with Kakao to bring ChatGPT responses into KakaoTalk group chats. The company said Korean organisations including LG Electronics, LG Uplus, LG CNS, GS E&C, Samsung SDS, TVING, Krafton, Toss, MUSINSA, Korea Zinc, Nexen Tire, and HanaTour are using ChatGPT Enterprise, OpenAI APIs, or Codex. (Photo by Zulfugar Karimov) See also: Omio scales travel product development using OpenAI models Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Samsung opens ChatGPT Enterprise and Codex access after AI restrictions appeared first on AI News. View the full article

Anthropic launched a beta version of its Claude Tag feature for Enterprise and Team tiers, shifting its chat model into shared Slack channels. Moving away from traditional isolated chat boxes, users pull the artificial intelligence model into active group threads by typing @Claude. The integration allows any team member in the channel to delegate a task, review the model’s outputs, and pick up the discussion thread from a previous point. This structural shift follows a US$65 billion Series H funding round that brought Anthropic’s post-money valuation to US$965 billion, positioned above rival OpenAI’s US$852 billion mark. Following a confidential S-1 filing for an initial public offering, market competition for business software placement remains tight. Data from corporate expense platform Ramp’s May 2026 AI Index indicates Anthropic’s enterprise adoption rate reached 34.4%, passing OpenAI’s 32.3% footprint. Modifying the channel workstream Standard generative software requires enterprise employees to move data between team chats and separate browser instances. Anthropic aims to reduce this back-and-forth movement by restructuring workplace AI agents to work in multiplayer environments. “Instead of a private back-and-forth, Claude Tag shows up in the open,” stated Rob Seaman, general manager of Slack, regarding the operational mechanics of the application. This shared visibility alters how context is tracked inside an organisation. Because Claude Tag logs its task status directly inside the communication window, multiple employees can monitor the live execution steps. The system tracks ongoing information from its active channels to build a contextual background. This automated history tracking limits the need for team members to continuously retype foundational company data or project scopes. Functional mechanics and asynchronous tasks The technical foundation for this channel integration relies on Anthropic’s Opus 4.8 engine. When assigned a request, the model divides the operation into sequential execution phases and utilises connected corporate databases, tools, and code repositories to complete the work. The primary operational difference for these workplace AI agents is their capability to function asynchronously without real-time human prompting. If a network administrator activates the tool’s “ambient” configuration, Claude Tag monitors threads and tracks tasks autonomously. The agent checks inactive text threads, signals priority notifications from integrated software extensions, and tracks unresolved assignments across multi-day intervals. Cat Wu, head of product for Claude Code, noted that the change centres on user configuration rather than completely new logic. “The form factor of being able to tag it the same way that you would a coworker is really powerful,” Wu told Reuters. Wu explained that connecting her personal Claude Tag agent to her email archive allows the system to analyse incoming communications, categorise urgent entries, and send immediate alerts inside Slack. Metrics and administrative controls Internal reporting from Anthropic shows that automated code generation has altered engineering activities, with the firm’sinternal product group creating 65% of its code through its private version of Claude Tag. Beyond software development, the vendor targets non-technical office workforces. Early customer implementations focus on querying database metrics, parsing analytics data, and processing internal IT support tickets. This expansion of background agent operations requires a distinct security infrastructure to protect proprietary information. To restrict data access to approved departments, system administrators must establish scoped Claude identities. All localised memories and tool integrations are confined strictly to specific channels authorised by the IT department. Additionally, management portals offer full tracking logs of user queries alongside specific organisational caps to regulate monthly token costs. The enterprise calculation: Autonomy vs. governance Frankly, moving generative tools from individual sandboxes into persistent corporate communication channels presents distinct operational trade-offs. The clear upside is the optimisation of routine knowledge work. By centralising information logs directly inside active threads, companies can lower task friction, capture context across changing project teams, and reduce the time spent on manual codebase tracking or database updates. However, delegating cross-app workflows to background agents introduces significant structural risks for IT departments. Permitting automated systems to read chat histories, connect to email accounts, and modify central code repositories expands an organisation’s internal data-exposure risks. If access boundaries are misconfigured, sensitive proprietary context could cross into unapproved channels. Furthermore, autonomous asynchronous execution removes direct human verification from intermediate workflow stages, leaving teams vulnerable to systemic errors if the underlying model misinterprets instructions mid-task. Corporate decision-makers must ultimately evaluate whether the productivity gains of channel-based automation outweigh the rigorous auditing, compliance overhead, and channel-by-channel security configurations required to safely govern an always-on agent. See also: Anthropic releases Claude Opus 4.8 Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Anthropic drops ‘workplace AI agents’ directly inside Slack appeared first on AI News. View the full article

Omio integrates OpenAI models across its engineering operations to accelerate travel product development and launch booking interfaces. The multimodal travel platform coordinates operations with over 3,000 transportation providers across 47 countries. Omio explicitly rejects the superficial addition of technology to outdated internal processes. The company’s CTO, Tomas Vocetka, requires all internal functions to completely redesign their operational execution frameworks from the ground up to operate as a native AI enterprise. OpenAI Codex integration Vocetka initiated the internal deployment by providing base ChatGPT access to the workforce, establishing a baseline familiarity with generative models before executing the primary technical integration. Omio subsequently embedded OpenAI Codex directly into its engineering operations, mandating its application across the entire software development lifecycle. Engineers currently apply Codex to preliminary research, architectural planning, active coding, automated testing, code reviews, and ongoing system maintenance. The engineering division constructs custom internal connectors to link proprietary data environments directly with these tools. This setup allows developers to bypass basic information retrieval and proceed directly to active task execution within their integrated development environments. Vocetka categorises the initial ChatGPT rollout as a preliminary introduction, emphasising that Codex handles the actual production workload. The deployment execution matured beyond the technical divisions. Management actively expands the use of Codex into non-technical corporate functions across the wider organisation. This expansion ensures standard operational procedures adapt to the new capabilities introduced by the engineering team. Internal analysis indicates the technical effort required to build specific products now sits at approximately 20 percent of previous levels. Delivery timelines show corresponding compression. Projects demanding the attention of multiple developers over an entire fiscal quarter now require a single engineer operating for roughly one month. Faster cycle times allow the engineering teams to test experimental concepts and validate consumer demand with minimal resource expenditure. Management allocates capital and engineering hours with greater precision, relying on prototyping to eliminate unviable features before committing to full-scale production. Lowering the time and cost barrier for software creation enables quicker internal decision-making. The technical teams iterate on existing products at a much higher velocity, pushing updates and new interface elements to the live environment at accelerated pace. Conversational commerce built on real-time transport data Omio launched one of the earliest conversational travel booking interfaces in 2023 by connecting OpenAI models to its proprietary transportation inventory. The system processes natural language queries regarding complex multimodal routes. Travelers input natural language requests asking for the fastest route from Rome to Florence, or comparing flights and trains between Paris and Barcelona. Omio aggregates services spanning trains, buses, ferries, and flights. Legacy travel booking required users to navigate multiple websites, manually compare modes of transport, and independently aggregate itineraries across multiple providers. Omio replaces this fractured process with a unified interface capable of parsing consumer intent. The generative models analyse text inputs and ping the booking systems to construct viable travel paths. The application functions by grounding the model responses in live pricing and availability data. The architecture prevents the generation of travel options based on static or outdated training data. The resulting output provides consumers with directly bookable itineraries. Omio expanded its initial integration into a dedicated ChatGPT experience. This dedicated application directly accesses the global transportation network maintained by the company. By grounding the user interaction in verified data, the technical team ensures high-fidelity responses. Consumers receive highly-personalised journey options rather than generic travel advice. Omio defines this structural setup as a new category of conversational commerce. The AI operates as the primary interface layer mediating the interaction between the consumer and the underlying global transportation network. The company views this as a broader departure from legacy search-based interfaces toward native generative customer experiences. The deployment points to a future where travel planning relies entirely on interacting with intelligent systems connected directly to live transportation networks. Omio’s corporate policy explicitly mandates that human personnel retain full accountability for all deployed code and final business outcomes. Generative tools function strictly as acceleration engines for development, analysis, and decision-making. “The responsibility and accountability stay with people. AI helps us develop faster, analyse faster, and make decisions faster, but people stay in charge,” explains Vocetka. This governance structure prevents automated systems from independently executing irreversible changes to the booking infrastructure or the core multimodal routing algorithms. The combination of broad employee access to OpenAI tools and rigorous oversight models creates an environment prioritising both speed and systemic stability. See also: Mitigating vendor lock-in with Sakana AI Fugu multi-agent models Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Omio scales travel product development using OpenAI models appeared first on AI News. View the full article

The global surge in AI cyber threats is no longer a distant problem for corporate data centres, according to an urgent public warning from the world’s most powerful intelligence alliance. On June 22, 2026, the cybersecurity chiefs of the Five Eyes nations—comprising the US, ***, Canada, Australia, and New Zealand—issued a rare joint intelligence briefing stating that upcoming artificial intelligence models will supercharge offensive hacking capabilities on a timeline measured in months, not years. While the advisory specifically tells corporate executives to overhaul their network defences, the rapid evolution of these tools means everyday internet users are about to face a much shiftier digital landscape. The massive shift in AI cyber threats The intelligence brief highlights an immediate danger: advanced, upcoming models like OpenAI’s “GPT-5.5-Cyber” and Anthropic’s “Mythos” are actively lowering the technical barriers for digital crime. Rogue actors no longer need elite coding skills to build complex, devastating software exploits. Instead, automated digital agents can scan internet-connected infrastructure around the clock to find software vulnerabilities before human engineers can patch them. This drastically shrinks the safety window that technology companies rely on to keep user applications secure. How does this hit home for regular users? When criminal networks use automated tools to breach large databases, the immediate consequence is the theft of regular consumer data. Your personal information, saved passwords, and cloud backups are the ultimate targets in these accelerated corporate intrusions. Furthermore, bad actors are leveraging conversational models to generate hyper-personalised phishing scams at an industrial scale. This trend is hitting the Asia-Pacific (APAC) region particularly hard, with countries like India recording a staggering 165% spike in ransomware incidents in early 2026 due to AI-assisted targeting. Rather than relying on easily spotted, poorly written spam emails, automated systems can scan your public social media profiles to write flawless, highly convincing messages designed to steal your credentials. Fighting back with the same tech The primary challenge facing cyber defenders is that machine-paced offence naturally moves faster than human-led detection. According to the World Economic Forum’s Global Cybersecurity Outlook, a massive 94% of corporate executives identify AI as their top threat vector, yet two out of three organisations report moderate to critical cybersecurity talent shortages. Network administrators are finding it impossible to review and deploy traditional security patches manually when rogue AI agents can discover and exploit a software vulnerability within minutes. The Five Eyes alliance emphasises that the most effective way to withstand these accelerating AI cyber threats is to deploy automated defences. Security teams are actively integrating defensive artificial intelligence models to monitor unusual behaviour and isolate network breaches. For individual users, the basic rules of internet safety are becoming mandatory. Turning on multi-factor authentication and deleting old, unused online accounts remain the most effective ways to break the automated chain of an AI-driven attack. See also: AI web search risks: Mitigating business data accuracy threats Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Top spy agencies say AI cyber threats will impact you within months. Here’s why appeared first on AI News. View the full article