ChatGPT

Open-source document database platform RavenDB has launched what it calls “the first fully integrated database-native AI Agent Creator,” a tool that makes it easier for enterprises to build and deploy AI agents. The platform tackles a common problem in enterprise AI – the difficulty of connecting models to a company’s own data systems and workflows securely and cost-effectively. Making AI practical, not just powerful The company wants to make AI deployment faster and more secure. Oren Eini, CEO and Founder of RavenDB, said the goal is to make AI deliver real value by embedding it directly where company data already lives. He explained that many organisations struggle because their data is scattered in multiple systems and formats, making integration expensive and complex. “The biggest problem users have with building AI solutions is that a generic model doesn’t actually do anything valuable,” he said. “For AI to bring real value into your system, you need to incorporate your own systems, data, and operations.” RavenDB’s new AI Agent Creator eliminates much of the overhead by letting companies expose relevant data to a model directly in the database – without separate vector stores or ETL workflows. The system manages technical challenges automatically, like model memory handling, summarisation, and data security. According to Eini, this means companies “can move from an idea to a deployed agent in a day or two.” Direct data access and real-time answers Traditional AI workflows usually involve exporting data from a database to a vector store, then connecting that store to an AI model, creating delays and security gaps. RavenDB’s approach uses built-in vector indexing and semantic search to make information available instantly to AI agents inside the database itself. That design supports real-time responsiveness, letting an AI agent access newly-updated information immediately: For example, checking a customer’s latest order or shipment status without waiting for a data refresh. On the question of security, Eini said: “An AI agent will not be executed as a privileged part of the system,” he noted. “It functions as an external entity with the same access rights as the user operating it.” Use cases and industry insight Eini noted that RavenDB has already applied the AI Agent Creator in real customer environments. In one example, the system is used for candidate ranking in recruitment, automatically reading and comparing uploaded resumés against job requirements to identify promising applicants. In another example, Eini explained how AI Agent Creator is being used to re-rank semantic search results to output accurate relevance rather than just find the nearest vector matches. Industry analysts see this kind of integration as part of a larger shift toward embedded, domain-specific AI. In a recent Forrester report, senior analyst Stephanie Liu wrote, “AI agents are eyeing autonomy, but your poor documentation means they may not reach this threshold.” She said that while full autonomy remains challenging, tighter links between AI systems and live enterprise data can “deliver immediate, practical value” for organisations experimenting with agentic AI. Broader context Database-native AI could mark a big shift in how companies use machine intelligence in their operations. By keeping both compute and security barriers inside the database, platforms like RavenDB could cut down on the need for additional infrastructure layers – a challenge many businesses face as they scale their AI programmes. AI News recently covered Google’s Gemini Enterprise, which aims to bring AI agents into everyday business workflows, and examined how CrateDB is rethinking database infrastructure for real-time AI performance. These are two major developments that reflect how agentic systems and data-centric architectures converge to make enterprise AI more efficient. RavenDB’s latest addition builds on that trend, positioning databases as active participants in AI pipelines, not passive data dumps. Looking ahead Eini said the launch reflects RavenDB’s roadmap to make AI capabilities a native part of its platform. Over the past year, the company has added vector search, embedding generation, and generative AI features directly into the database engine. “We aim to encapsulate all the AI complexity inside RavenDB,” he said, “so users can focus on the results rather than the mechanics.” As enterprises continue to seek reliable, cost-efficient ways to adopt AI, database-native tools like RavenDB’s AI Agent Creator may offer a practical path forward, merging operational data and intelligence in one environment. Image source: Unslpash The post RavenDB launches database-native AI agent creator to simplify enterprise AI integration appeared first on AI News. View the full article

With AI workloads becoming more computationally demanding, organisations across the globe are fast realising that traditional centralised providers aren’t always the answer to their burgeoning needs. And while compute giants (like AWS, Google Cloud, and Azure) continue to capture the limelight when it comes to AI processing, a quieter revolution has been brewing. Enter the decentralised compute marketplace, consisting of platforms capable of connecting organisations needing GPU power with providers who have hardware to spare, via decentralised mechanisms. These marketplaces are capable of handling real workloads like AI model training and 3D rendering at costs that traditional cloud providers can’t compete with. In this article, we will discuss five platforms worth watching as they strive to reshape how computing resources are allocated. 1. Argentum AI Building a liquid marketplace for computing is harder than it sounds, yet Argentum AI has tackled this challenge successfully by treating GPU resources the way financial markets treat stocks. That is, as tradable commodities with transparent pricing and real-time settlement. The platform operates as an independent, decentralised marketplace where enterprises can post computing tasks and providers bid to execute them. However, what truly sets the platform apart is its underlying infrastructure. Argentum uses real-time bidding, verifiable execution, and blockchain-based transparent settlement to unlock idle computing capacity. Settlement happens on-chain, with Ethereum smart contracts holding funds in escrow until jobs complete successfully. The power of this model was put on full display when, earlier in October, Argentum closed an oversubscribed pre-seed funding round led by Kraken, Banyan Ventures, Victor Morganstern, and Todd Bensen. The platform plans to work in conjunction with GPU manufacturers to establish liquidity and monetisation plans for second-life assets, reducing its compute costs even further. 2. Aethir Scale matters in the infrastructure realm, and Aethir has achieved the feat quickly. Within a year (following its Token Generation Event), the project has established itself as one of the largest decentralised GPU clouds in today’s Web3 economy. The platform sources GPU capacity from tier 3 and tier 4 data centres, making them available through a distributed network of 3,000+ NVIDIA H100s and H200s, plus 62,000+ Aethir Edge cloud computing devices.To allay any pricing volatility-related concerns (especially for those looking to pay the firm in cryptocurrency), Aethir partnered recently with Maitrix to introduce AUSD, an algorithmic stablecoin pegged to the US dollar. 3. Bittensor If decentralised compute marketplaces are disruptive, Bittensor takes it to the next level by turning AI itself into a marketplace. It does so by operating as an L1, where developers can train AI models and contribute machine intelligence in lieu of the network’s native TAO token. Earlier this year, the Bittensor team introduced Dynamic TAO (dTAO), an upgrade that allows each subnet to issue its own Alpha Token and compete for TAO rewards through open market mechanisms (indirectly creating a competitive environment where the best AI models and subnets naturally attract more resources). 4. Akash Network Akash takes a straightforward approach to decentralised cloud operations, in that it matches idle computing resources with flexible demand through an open marketplace. The platform allows users to rent computing resources from a global network of providers, with costs up to 80% lower than traditional cloud services’. The network runs on Cosmos SDK and uses a Delegated Proof-of-Stake consensus mechanism where users can specify their exact requirements (like CPU, memory, storage, geographic location) and providers can bid for these requests. In August 2025, Akash partnered with NVIDIA to deploy Blackwell B200/B300 GPUs on its decentralised cloud, targeting AI developers needing high-performance training and inference. 5. Flux Flux combines the power of the blockchain with cloud computing through a unique Proof-of-Useful-Work v2 model, which replaces GPU mining with a node-centric system where FluxNodes running real workloads secure the network (reducing emissions by 10% annually and targeting sub-1% inflation). The platform encompasses FluxOS (a Linux-based OS for deploying decentralised apps), FluxEdge (a GPU rental platform for AI/ML workloads), and Zelcore (a multi-chain wallet supporting 85+ blockchains). The post Top 5 AI compute marketplaces reshaping the landscape in 2026 appeared first on AI News. View the full article

Although people use AI extensively for personal and business purposes, they don’t necessarily fully trust it. For business leaders, using AI is no longer just an option but something they must do to stay competitive. Putting AI into everything, from helpful assistants to automatic processes, could boost productivity and open up new ways to make money. CIOs and Chief Data Officers have the difficult job of guiding their companies through digital changes. Even though workers and customers are using AI tools, they don’t really trust them, which brings up new worries about how to control things and could hurt a company’s reputation. Plenty of use, but not enough trust In the UAE – which is known for quickly adopting new technology – 97 percent of people are using AI for their jobs, studies, or personal stuff, according to a report from KPMG. That’s one of the highest rates in the world, but this widespread use hides some deep worries. The same report said that 84 percent of people would only trust AI systems if they were sure they were being used in a trustworthy way, and 57 percent think there need to be stronger rules to make AI feel safe. Clearly, even though people are using AI a lot, trust hasn’t caught up. And it’s not just happening in the UAE. Information from the *** shows a similar, maybe more developed, gap in trust. KPMG found that only 42 percent of people in the *** are willing to trust AI. While 57 percent accept or approve of using it, 80 percent think stronger rules are needed to make sure it’s used responsibly. These AI trust numbers should worry business leaders. 78 percent of people in the *** are concerned about bad things that could happen because of AI, and only 10 percent say they know about the AI rules that already exist in the country. The information suggests that even in places where technology is common, trust and understanding are still way behind how much AI is being used. When 80 percent of an important market wants stronger rules, it’s a sign to the people in charge that they’re not doing a good enough job of keeping things in check. Putting a new AI customer service tool into a market that already doesn’t fully trust AI is a big risk to a company’s image. The difference between using and trusting AI will be a key point in the next stage of digital change, according to Lei Gao, CTO at SleekFlow. Lei says that using AI isn’t the problem anymore, but being responsible is. People are okay with using AI as long as they think it’s being used in a responsible way. “Adoption is no longer the issue; accountability is. People are comfortable using AI as long as they believe it’s being used responsibly,” explains Lei. “In customer communication, for example, users trust AI when it behaves predictably and transparently. If they can’t tell when automation is making a decision, or if it feels inconsistent, that trust starts to erode.” For companies working in markets where AI is used a lot but not trusted, the most important thing should be to create trust. Lei believes that leaders need to include openness, consistency, and human control in every AI interaction. This means having clear rules, whether a company is using basic AI models from AWS Bedrock, managing information in a Dell AI Factory, or using AI helpers like SAP Joule. A plan to build AI trust in your business To fix this problem, Lei suggests three main ideas that turn AI strategy from a technical issue into a question of control: Companies need to be open about when AI is being used. Customers and workers appreciate honesty. Lei suggests making it clear when people are talking to AI and when a human takes over. This simple action is essential because clarity builds trust. Technology leaders need to use AI to help people, not leave them out. This is important for getting people to use AI within the company and reducing resistance. Lei says that AI should make people better, not replace them. Business leaders need to check AI for its tone and fairness. This is where control becomes an ongoing process. Lei notes that using AI responsibly is something that needs to be done all the time; it’s not enough to just launch it and forget about it. Keeping trust and following the rules means regularly checking the tone, bias, and how well AI handles problems. The UAE has proven that AI can be adopted faster than almost anywhere else, but how fast it’s being put in place isn’t the best measure of success anymore. The challenge for business leaders is to show that their AI systems are not just strong, but also reliable, fair, open, and can earn trust. “The next milestone is trust and showing that automation can work in the service of people, not just performance metrics,” Lei concludes. See also: OpenAI connects ChatGPT to enterprise data to surface knowledge 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 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 fix the AI trust gap in your business appeared first on AI News. View the full article

Huawei has released its CloudMatrix 384 AI chip cluster, a new system for AI learning. It employs clusters of Ascend 910C processors, joined via optical links. The distributed architecture means the system can outperform traditional hardware GPU setups, particularly in terms of resource use and on-chip time, despite the individual Ascend chips being less powerful than those of competitors. Huawei’s new framework positions the tech giant as a “formidable challenger to Nvidia’s market-leading position, despite ongoing US sanctions,” the company claims. To use the new Huawei framework for AI, data engineers will need to adapt their workflows, using frameworks that support Huawei’s Ascend processors, such MindSpore, which are available from Huawei and its partners Framework transition: From PyTorch/TensorFlow to MindSpore Unlike NVIDIA’S ecosystem, which predominantly uses frameworks like PyTorch and TensorFlow (engineered to take full advantage of CUDA), Huawei’s Ascend processors perform best when used with MindSpore, a deep learning framework developed by the company. If data engineers already have models built in PyTorch or TensorFlow, they will likely need to convert models to the MindSpore format or retrain them using the MindSpore API. It is worth noting that MindSpore uses different syntax, training pipelines and function calls from PyTorch or TensorFlow, so a degree of re-engineering will be necessary to replicate the results from model architectures and training pipelines. For instance, individual operator behaviour varies, such as padding modes in convolution and pooling layers. There are also differences in default weight initialisation methods. Using MindIR for model deployment MindSpore employs MindIR (MindSpore Intermediate Representation), a close analogue to Nvidia NIM. According to MindSpore’s official documentation, once a model has been trained in MindSpore, it can be exported using the mindspore.export utility, which converts the trained network into the MindIR format. Detailed by DeepWiki’s guide, deploying a model for inference typically involves loading the exported MindIR model and then running predictions using MindSpore’s inference APIs for Ascend chips, which handle model de-serialisation, allocation, and execution. MindSpore separates training and inference logic more explicitly than PyTorch or TensorFlow. Therefore, all preprocessing needs to match training inputs, and static graph execution must be optimised. MindSpore Lite or Ascend Model Zoo are recommended for additional hardware-specific tuning. Adapting to CANN (Compute Architecture for Neural Networks) Huawei’s CANN features a set of tools and libraries tailored for Ascend software, paralleling NVIDIA’s CUDA in functionality. Huawei recommends using CANN’s profiling and debugging tools to monitor and improve model performance on Ascend hardware. Execution Modes: GRAPH_MODE vs.PYNATIVE_MODE MindSpore provides two execution modes: GRAPH_MODE – Compiles the computation graph before execution. This can result in faster execution and better performance optimisation since the graph can be analysed during compilation. PYNATIVE_MODE – Immediately executes operations, resulting in simpler debugging processes, better suited, therefore, for the early stages of model development, due to its more granular error tracking. For initial development, PYNATIVE_MODE is recommended for simpler iterative testing and debugging. When models are ready to be deployed, switching to GRAPH_MODE can help achieve maximum efficiency on Ascend hardware. Switching between modes lets engineering teams balance development flexibility with deployment performance. Code should be adjusted for each mode. For instance, when in GRAPH_MODE, it’s best to avoid Python-native control flow where possible. Deployment environment: Huawei ModelArts As you might expect, Huawei’s ModelArts, the company’s cloud-based AI development and deployment platform, is tightly integrated with Huawei’s Ascend hardware and the MindSpore framework. While it is comparable to platforms like AWS SageMaker and Google Vertex AI, it is optimised for Huawei’s AI processors. Huawei says ModelArts supports the full pipeline from data labelling and preprocessing to model training, deployment, and monitoring. Each stage of the pipeline is available via API or the web interface. In summary Adapting to MindSpore and CANN may necessitate training and time, particularly for teams accustomed to NVIDIA’s ecosystem, with data engineers needing to understand various new processes. These include how CANN handles model compilation and optimisation for Ascend hardware, adjusting tooling and automation pipelines designed initially for NVIDIA GPUs, and learning new APIs and workflows specific to MindSpore. Although Huawei’s tools are evolving, they lack the maturity, stability, and broader ecosystem support that frameworks like PyTorch with CUDA offer. However, Huawei hopes that migrating to its processes and infrastructure will pay off in terms of results, and let organisations reduce reliance on US-based Nvidia. Huawei’s Ascend processors may be powerful and designed for AI workloads, but they have only limited distribution in some countries. Teams outside Huawei’s core markets may struggle to test or deploy models on Ascend hardware, unless they use partner platforms, like ModelArts, that offer remote access. Fortunately, Huawei provides extensive migration guides, support, and resources to support any transition. (Image source: “Huawei P9” by 405 Mi16 is licensed under CC BY-NC-ND 2.0.) 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 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 Re-engineering for better results: The Huawei AI stack appeared first on AI News. View the full article

OpenAI is surfacing company knowledge by connecting ChatGPT to enterprise data, turning it from a general assistant into a custom analyst. For business leaders, generative AI’s potential has always been limited by its lack of access to internal data. Even the best AI isn’t helpful if it can’t access the info needed to do a job. OpenAI points out that the info you need is often in your internal tools, but that knowledge is scattered across documents, files, messages, emails, tickets, and project trackers. This scattering is more than just annoying; it hurts efficiency and decisionmaking. The main problem is that these tools don’t always connect, and the best answer is often spread across all of them. This puts OpenAI up against the AI strategies of big enterprise platforms like Microsoft’s Copilot in Azure and Office 365, Google’s Vertex AI, Salesforce’s Agentforce, and AWS Bedrock. Everyone is racing to connect models to secure company data. OpenAI uses third-party data for ChatGPT enterprise tasks ChatGPT will connect to apps like Slack, SharePoint, Google Drive, and GitHub. OpenAI says it’s powered by a version of GPT-5, trained to check many sources for better answers. For checking and validation, every answer shows where the info came from. This changes what you can do from simple writing to complex analysis. For example, a manager prepping for a client call could ask for a briefing. The model could then use recent Slack messages, email details, call notes from Google Docs, and support tickets from Intercom to make a summary. This power can also handle confusion. If you ask, “What are the company goals for next year?”, the tool will summarise what’s been talked about and point out different opinions. This goes beyond just finding data; now it’s analysing situations and helping leaders find disagreements or unfinished decisions. Other uses for teams: Strategy: Putting together customer feedback from Slack, survey results from Google Slides, and main topics from support tickets to plan roadmaps. Reporting: Making campaign summaries by getting data from HubSpot, briefs from Google Docs, and key points from emails. Planning: Helping engineering leads plan releases by checking GitHub for open tasks, checking Linear for tickets, and checking Slack for bug reports. Addressing enterprise AI governance and implementation For CISOs and data leaders, sharing intellectual property with an AI model is a big risk. OpenAI is dealing with this by focusing on admin controls and data privacy. The most important control is that the system respects your current company permissions. OpenAI has ensured that ChatGPT can only see the enterprise data that each user can already see. ChatGPT Enterprise and Edu admins can manage access to apps and create custom roles. OpenAI says it doesn’t train on your data by default. It also has security features like encryption, SSO, SCIM, IP whitelisting, and a Compliance API for logs. But, tech leaders should know the limits. It’s not perfect yet. Users have to pick it when starting a conversation. Also, there’s a trade-off: when company knowledge is on, ChatGPT can’t search the web or make charts. OpenAI is working to fix this soon. The tool’s usefulness depends on its ecosystem. It’s launching with key platforms and adding connectors for tools like Asana, GitLab Issues, and ClickUp, copying the strategies of IBM watsonx and SAP Joule. OpenAI’s enterprise data knowledge surfacing is the next step for AI assistants like ChatGPT, moving them into the private core of businesses. It tries to solve the AI problem: connecting models to the data where work happens. For business leaders, this means: Check your data: Before using this, CISOs and CDAOs must check that data permissions in SharePoint, Google Drive, etc., are correct. The AI will only respect those permissions, so if they’re too open, the AI will show that weakness. Pilot with tricky tasks: Instead of rolling it out to everyone, find specific workflows that are slowed down by scattered info. Preparing client briefings or making cross-department reports are good places to start measuring results. Set expectations: Teams must know the limits. Having to manually turn it on and not being able to search the web at the same time are big limits to consider. Watch the ecosystem: The tool’s value will depend on its integrations. CIOs should compare the tool’s connector list to their company’s tech. Compare to current platforms: See how this compares to the AI solutions from Microsoft, Google, and Salesforce. The decision is quickly becoming about which data ecosystem offers the most secure, integrated, and cost-effective path. OpenAI’s new company knowledge feature shows that the most important thing for generative AI is now secure and useful data integration, not just how good the model is. This latest ChatGPT feature should make things much faster by getting rid of enterprise knowledge silos, but it also makes data governance and access control more important than ever. For business leaders, this tech isn’t a simple fix. Instead, it’s a good reason to get their data organised before others do. See also: OpenAI data residency advances enterprise AI governance 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 Expo, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post OpenAI connects ChatGPT to enterprise data to surface knowledge appeared first on AI News. View the full article

Anthropic’s announcement this week that it will deploy up to one million Google Cloud TPUs in a deal worth tens of billions of dollars marks a significant recalibration in enterprise AI infrastructure strategy. The expansion, expected to bring over a gigawatt of capacity online in 2026, represents one of the largest single commitments to specialised AI accelerators by any foundation model provider—and offers enterprise leaders critical insights into the evolving economics and architecture decisions shaping production AI deployments. The move is particularly notable for its timing and scale. Anthropic now serves more than 300,000 business customers, with large accounts—defined as those representing over US$100,000 in annual run-rate revenue—growing nearly sevenfold in the past year. This customer growth trajectory, concentrated among Fortune 500 companies and AI-native startups, suggests that Claude’s adoption in enterprise environments is accelerating beyond early experimentation phases into production-grade implementations where infrastructure reliability, cost management, and performance consistency become non-negotiable. The multi-cloud calculus What distinguishes this announcement from typical vendor partnerships is Anthropic’s explicit articulation of a diversified compute strategy. The company operates across three distinct chip platforms: Google’s TPUs, Amazon’s Trainium, and NVIDIA’s GPUs. CFO Krishna Rao emphasised that Amazon remains the primary training partner and cloud provider, with ongoing work on Project Rainier—a massive compute cluster spanning hundreds of thousands of AI chips across multiple US data centres. For enterprise technology leaders evaluating their own AI infrastructure roadmaps, this multi-platform approach warrants attention. It reflects a pragmatic recognition that no single accelerator architecture or cloud ecosystem optimally serves all workloads. Training large language models, fine-tuning for domain-specific applications, serving inference at scale, and conducting alignment research each present different computational profiles, cost structures, and latency requirements. The strategic implication for CTOs and CIOs is clear: vendor lock-in at the infrastructure layer carries increasing risk as AI workloads mature. Organisations building long-term AI capabilities should evaluate how model providers’ own architectural choices—and their ability to port workloads across platforms—translate into flexibility, pricing leverage, and continuity assurance for enterprise customers. Price-performance and the economics of scale Google Cloud CEO Thomas Kurian attributed Anthropic’s expanded TPU commitment to “strong price-performance and efficiency” demonstrated over several years. While specific benchmark comparisons remain proprietary, the economics underlying this choice matter significantly for enterprise AI budgeting. TPUs, purpose-built for tensor operations central to neural network computation, typically offer advantages in throughput and energy efficiency for specific model architectures compared to general-purpose GPUs. The announcement’s reference to “over a gigawatt of capacity” is instructive: power consumption and cooling infrastructure increasingly constrain AI deployment at scale. For enterprises operating on-premises AI infrastructure or negotiating colocation agreements, understanding the total cost of ownership—including facilities, power, and operational overhead—becomes as critical as raw compute pricing. The seventh-generation TPU, codenamed Ironwood and referenced in the announcement, represents Google’s latest iteration in AI accelerator design. While technical specifications remain limited in public documentation, the maturity of Google’s AI accelerator portfolio—developed over nearly a decade—provides a counterpoint to enterprises evaluating newer entrants in the AI chip market. Proven production history, extensive tooling integration, and supply chain stability carry weight in enterprise procurement decisions where continuity risk can derail multi-year AI initiatives. Implications for enterprise AI strategy Several strategic considerations emerge from Anthropic’s infrastructure expansion for enterprise leaders planning their own AI investments: Capacity planning and vendor relationships: The scale of this commitment—tens of billions of dollars—illustrates the capital intensity required to serve enterprise AI demand at production scale. Organisations relying on foundation model APIs should assess their providers’ capacity roadmaps and diversification strategies to mitigate service availability risks during demand spikes or geopolitical supply chain disruptions. Alignment and safety testing at scale: Anthropic explicitly connects this expanded infrastructure to “more thorough testing, alignment research, and responsible deployment.” For enterprises in regulated industries—financial services, healthcare, government contracting—the computational resources dedicated to safety and alignment directly impact model reliability and compliance posture. Procurement conversations should address not just model performance metrics, but the testing and validation infrastructure supporting responsible deployment. Integration with enterprise AI ecosystems: While this announcement focuses on Google Cloud infrastructure, enterprise AI implementations increasingly span multiple platforms. Organisations using AWS Bedrock, Azure AI Foundry, or other model orchestration layers must understand how foundation model providers’ infrastructure choicesaffect API performance, regional availability, and compliance certifications across different cloud environments. The competitive landscape: Anthropic’s aggressive infrastructure expansion occurs against intensifying competition from OpenAI, Meta, and other well-capitalised model providers. For enterprise buyers, this capital deployment race translates into continuous model capability improvements—but also potential pricing pressure, vendor consolidation, and shifting partnership dynamics that require active vendor management strategies. The broader context for this announcement includes growing enterprise scrutiny of AI infrastructure costs. As organisations move from pilot projects to production deployments, infrastructure efficiency directly impacts AI ROI. Anthropic’s choice to diversify across TPUs, Trainium, and GPUs—rather than standardising on a single platform—suggests that no dominant architecture has emerged for all enterprise AI workloads. Technology leaders should resist premature standardisation and maintain architectural optionality as the market continues to evolve rapidly. See also: Anthropic details its AI safety strategy 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 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’s billion-dollar TPU expansion signals a strategic shift in enterprise AI infrastructure appeared first on AI News. View the full article

Lightricks is upping the ante for rapid video creation and iteration with its latest artificial intelligence model. The company claims its newly released LTX-2 foundation model can generate new content faster than playback speed, plus it raises the bar in resolution and quality. The open-source LTX-2 can generate a stylised, high-definition, six-second video in just five seconds without no compromise in quality, enabling creators to pump out professional content much faster than previously. It’s an impressive achievement, but it’s not the only parameter that sets LTX-2 apart from others. It combines native audio and video synthesis with open-source transparency, and if users are willing to wait just a few seconds longer, they can enhance their outputs to 4K resolution at up to 48 frames per second, the company says. Even better, creators can run the software on consumer-grade GPUs, dramatically reducing their compute costs. Diffusion models come of age LTX-2 is what’s known as a diffusion model, which works by incrementally adding “noise” to generated content and then reducing that noise until the output resembles the video assets the model has been trained on. With LTX-2, Lightricks has accelerated the diffusion process, so creators can iterate on their ideas by outputting live previews almost instantaneously. The model is also capable of generating accompanying audio at the same time – be it a soundtrack, dialogue or ambient sound effects – dramatically accelerating creative workflows. That’s a big deal, as before, creators would have had to conjure up any audio separately from the video, then spend time stitching it together and making sure there’s perfect synchronisation. Google’s Veo models have been celebrated for their powerful integration of synced sound generation, so these new capabilities in LTX serve to reinforce the idea that Lightricks’ tech is on par with the bleeding edge. When it comes to access options, Lightricks still offers creators plenty of flexibility with LTX-2. The company’s flagship LTX Studio platform is aimed at professionals, who, in some cases, are willing to sacrifice some speed to create videos at the highest quality. With the ensuing slightly slower rates of processing, they’ll be able to output videos in native 4K resolution at up to 48fps, creating at the same standard expected from cinematic productions, Lightricks claims. The platform offers a wide range of creative controls, affecting the model’s customisable parameters. More details on these will be announced soon, but should include pose and depth controls, video-to-video generation, and rendering alternatives – keep an eye out for a release date, later this autumn. Lightricks co-founder and Chief Executive Zeev Farbman believes that LTX-2’s enhanced capabilities illustrate the extent to which diffusion models are finally coming of age. He said in a statement that LTX-2 is: “The most complete and comprehensive creative AI engine we’ve ever built, combining synchronised audio and video, 4K fidelity, flexible workflows, and radical efficiency.” “The isn’t vaporware or a research demo,” he said. “It’s a real breakthrough in video generation.” A major milestone With LTX-2, Lightricks is demonstrating it’s at the cutting edge of AI video generation, with the platform coming on the back of a number of industry firsts in previous LTXV models. In July, the company’s family of LTXV models, including LTXV-2B and LTXV-13B, became the first to support long-form video generation, which followed an update extending output to up to 60 seconds. With this, AI video production became “truly directed,” with users able to start with an initial prompt, and add further prompts in real-time as video was being streamed live. LTXV-13B already had a reputation for being one of the most powerful video creation models around, even before that one minute update. Launching in May, it was the first platform in the industry to support multi-scale rendering, which let users progressively enhance their videos by prompting the model to add more colour and detail, step-by-step, in the same way that professional animators “layer” additional details on top of their work in traditional production processes. The 13B model was trained on licensed data from Getty and Shutterstock. The company’s partnerships with these content behemoths are important, not only for the quality of the training data, but also for ethical reasons; models’ outputs are far less problematic in terms of copyright, an issue that plagues many other AI models’ creations. Lightricks has also released a distilled version of LTXV-13B that simplifies and speeds up the diffusion process, meaning content can be generated in as little as four-to-eight steps. The distilled version also supports LoRAs, meaning it can be fine-tuned by users to create content that’s more attuned to the aesthetic style of a project. Innovative billing models Like those earlier models, LTX-2 will be released under an open-source licence, making it a viable alternative to Alibaba’s Wan2 series of models. Lightricks has stressed that it’s truly open-source, as opposed to just “open access,” which means that its pre-trained weights, datasets, and all tooling will be available on GitHub, alongside the model itself. LTX-2 is available to users in LTX Studio and through its API as of now, with the open-source version due to be released in November. For those who prefer to use the paid version via API, Lightricks offers flexible pricing, with costs starting at just $0.04 per second for a version that generates HD videos in just five seconds. The Pro version balances speed with performance, and here, prices start at $0.07 per second. The Ultra version costs $0.12 per second for video generation in 4K resolution at 48 fps, plus full-fidelity audio. Prices also vary according to resolution, with users able to choose between 720p, 1080p, 2K and 4K. Lightricks claims that thanks to the efficiency of the model’s processing, its pricing makes LTX-2 up to 50% cheaper than competing models, making extended projects more economically viable, yet with faster iteration and higher quality than previous generations. Alternatively, users will be able to use the model by downloading the open-source version and running it on consumer-grade GPUs after it lands on GitHub next month. Image source: Unsplash The post Open-source AI video from Lightricks offers 4K, sound, and faster rendering appeared first on AI News. View the full article

At its London Symbiosis 4 event on 22 October, Druid AI introduced what it terms Virtual Authoring Teams – a new generation of AI agents that can design, test, and deploy other AI agents. The announcement marks a move towards what the company calls a ‘factory model’ for AI automation. According to Druid, the system enables organisations to build enterprise-grade AI agents up to ten times faster, and the platform offers orchestration facilities, plus compliance safeguards and measurable ROI tracking. The orchestration engine, Druid Conductor, serves as a control layer that integrates data, tooling, and human oversight into a single framework. In addition to the Druid Conductor is the Druid Agentic Marketplace, a repository of pre-built, industry-specific agents for banking, healthcare, education, and insurance. With its solutions, Druid wants to make agentic AI accessible to non-technical users, but provide scalability capability suitable for enterprise use. Chief Executive Joe Kim described it as “AI [that] actually works” – a bold claim in a market flooded with experimentation and unproven automation frameworks. The new agentic battleground Druid is not alone in its pursuit. Similar platforms, the likes of Cognigy, Kore.ai, and Amelia, each represent heavy investment in multi-agent orchestration environments. OpenAI’s GPTs and Anthropic’s Claude Projects also allow users to design semi-autonomous digital workers without coding expertise. Google’s Vertex AI Agents and Microsoft’s Copilot Studio are moving in the same direction, placing agentic AI as an extension to enterprise ecosystems rather than stand-alone products. The difference between the competing platforms lies in execution – some focus on workflow automation, others on conversational depth or ease of integration with other parts of the IT stack. For technology buyers, such diversity is an opportunity and a risk. Vendors are racing to define what agentic AI means in practice, and there’s an undoubted element of agentic AI being 2025’s buzzword, implying differentiation between pure LLM models and practical tools useful in business contexts. Some vendors view agentic as an architecture – modular, distributed, and explainable, while others frame agentic AI as a layer of automation that builds itself – or rather, can discover what powers it’s been granted, and use them according to natural language instructions. The truth of agentic AI’s abilities sits somewhere between engineering promises and operational reality. The business case – and the caveats Agentic AI systems promise extraordinary benefits. They can accelerate routine development, coordinate multiple business functions, and use data repositories that were once siloed. For enterprises under pressure to deliver digital transformation with limited headcount, the idea of self-building AI teams is compelling. But the use of the conditional tense in many vendors’ marketing materials and descriptions is telling: agentic AI can achieve savings, could drive faster operations, and so on. Business leaders should approach such systems with a clear head. There are few proven case studies beyond pilot programmes inside large corporations (those with mature data governance and deep budgets), and even in those organisations, the returns have been uneven. Failures are rarely shouted from the rooftops, after all. The biggest risks are not technical – they’re organisational. Delegating complex decision-making to automated agents without sufficient oversight introduces potential bias, compliance breaches, and reputational exposure. Systems can also generate automation debt: a growing tangle of interconnected bots that become difficult to monitor or update as business processes evolve. The issue of necessary organisational change is troubling on two counts, furthermore. Most business processes have evolved a particular way for good reasons, so why change them to implement a new, largely unproven technology? Secondly, what’s often proposed is change that’s instigated by technology implementation. Shouldn’t processes change for strategic reasons, and technology support that change? Is this a case of the IT tail wagging the business dog? Security remains a further concern. Each agent increases the surface area for potential breaches or data misuse, particularly when they are designed to communicate and collaborate autonomously. As more workflows become self-directed, ensuring traceability and accountability becomes essential, and more difficult to unpick as complexity increases. The necessary headcount to monitor results and ensure rigorous oversight could negate any ROI agentic AI offers. Why agentic AI attracts enterprises Despite the challenges, the attraction is easy to understand. A successful agentic system can transform the speed at which an enterprise experiments and scales. By delegating repeatable cognitive tasks – from compliance checks to customer service triage – organisations can redirect human activity elsewhere. Druid’s Virtual Authoring Teams encapsulate the logic: automate the automation. Its marketplace of domain-specific agents offers enterprises a head start, promising faster deployments and measurable ROI. For sectors struggling with talent shortages and regulatory pressure, that is an appealing prospect. Moreover, Druid’s emphasis on explainable AI and its orchestration layer suggests an awareness of corporate caution. Its stated pillars – control, accuracy, and results – are designed to reassure boards that transparency can coexist with speed. If the system truly delivers what the company claims, it could narrow the gap between AI experimentation and scalable transformation. Balancing autonomy with accountability Still, for every organisation embracing agentic AI, another remains unconvinced. Many enterprises are wary of over-promising vendors and pilot fatigue. A technology capable of designing and deploying its own successors raises operational questions. What happens when an agent acts beyond its creator’s intent? How do governance frameworks keep pace? Business leaders must treat autonomy as a spectrum, not a goal. The near future of enterprise AI will likely blend human-supervised automation with limited agentic autonomy. Systems like Druid’s may act as orchestration hubs rather than fully independent actors. From hype to utility Agentic AI represents a natural evolution of automation in a wild frontier. Its potential is obvious, yet the market still lacks broad, evidence-based validation of sustained business outcomes. It may just be early days, or may be hyperbole drowning out the voices of reason. For now, agentic systems do work in controlled contexts – contact-centre operations, document processing, and IT service management. Scaling agentic AI across organisations will require maturity not just in technology, but in culture, process design, and methods of oversight. As Druid and its peers expand their offerings, enterprises will need to weigh the cost of control against the promised wins from better automation. The next two years will determine whether AI factories become a part of business operations, or another layer of abstraction with its own overheads. (Image source: “****** and grey wolf (female from Druid pack,’Half ******’) walking in road near Lamar River bridge” by YellowstoneNPS is marked with Public Domain Mark 1.0. ) 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 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 Autonomy in the real world? Druid AI unveils AI agent ‘factory’ appeared first on AI News. View the full article

The emergence of artificial intelligence has fundamentally altered the field of content creation. Tools capable of generating coherent, often impressive, text, are now ubiquitous. Yet, despite their sophistication, AI-generated content presents a persistent challenge because it often has a “robotic” quality, lacking the warmth, nuance, and genuine voice that connects with a human audience. The digital coldness necessitates an important step: humanisation. This need has given rise to specialised AI humaniser tools designed to inject naturalness into machine prose. Simultaneously, the age-old craft of human editing continues to refine and elevate text. This brings us to a pivotal question for anyone producing written content today. When transforming raw AI output into compelling prose, is it better to use an AI humaniser or rely on the irreplaceable touch of a human editor? The mechanics of the AI language humanisers An AI language humaniser is a sophisticated software application engineered to transform text produced by large language models into prose that more closely resembles human authorship. Its primary objective is to smooth out the typical AI “tells,” like repetitive sentence structures, overly formal or generic vocabulary, and a lack of emotional depth or personal voice, making the content sound more natural and engaging. The tools use advanced algorithms and machine learning to analyse text and modify it for characteristics like perplexity (how predictable the text is) and burstiness (the variation in sentence length and structure), aiming to make the output less uniform. For instance, you can humanise AI with StudyAgent to make academic essays read more coherently, adapt business reports to a specific brand voice, or fine-tune SEO articles to be more engaging for readers, and still retain keyword density. They serve as a bridge between raw machine output and audience-ready content in various domains. The artistry of human editorial craft In stark contrast to algorithmic processing, human editors bring an unparalleled depth of understanding and artistry to text. Their work goes far beyond surface-level corrections, delving into the essence of communication. An experienced human editor handles the intricate dance of style, tone, and contextual relevance, ensuring the message resonates deeply with its intended audience. The unique strengths of human editing are numerous and qualitative: Creative infusion: Editors can inject genuine creativity, metaphorical language, and evocative descriptions that AI often struggles to generate authentically. Intuitive grasp of nuance: They possess an innate intuition for subtle meanings, subtext, and implied emotion, adapting text to achieve specific psychological or rhetorical effects. Cultural and social sensitivity: Human editors navigate complex cultural nuances, local idioms, and social sensitivities that AI might misinterpret or overlook, preventing miscommunication. Voice and persona enhancement: They meticulously cultivate and amplify a unique authorial voice or brand persona, ensuring consistency and distinctiveness in all communications. Deep contextual insight: Editors can understand the broader strategic goals behind a piece of writing, tailoring it to achieve specific business, academic, or personal objectives. Refining text: Algorithmic versus humanistic approaches When we place AI language humanisers side-by-side with human editing, distinct differences emerge in several important aspects of content refinement. Each approach offers unique advantages depending on the specific demands of the project. Velocity of output AI humanisers: Offer near-instantaneous text transformation. You paste the content, select desired parameters, and receive a refined version almost immediately. This makes them invaluable for high-volume, rapid turnaround tasks. Human editing: Requires a longer timeframe. The process involves careful reading, thinking, multiple passes, and often communication between editor and author. This extended timeline is a trade-off for deeper quality. Stylistic elevation AI humanisers: Excel at smoothing out awkward phrasing, varying sentence structure, and replacing generic words with more diverse synonyms. They make AI text less robotic but may not instill a truly unique voice or personality. Human editing: Possesses the capacity to truly enhance an author’s distinct voice, inject personality, and ensure the tone is perfectly aligned with the intended impact. Editors can craft prose that is not just correct but compelling and memorable. Investment required AI humanisers: Represent a significantly lower financial investment and are typically available through subscription models or one-time fees that are a fraction of human editorial costs, making them accessible for budget-conscious users. Human editing: Is a premium service reflecting the skill, experience, and time investment of a professional. Costs are higher but justified by the depth of insight and bespoke quality delivered. Consistency of refinement AI Humanisers: Offer unwavering consistency. Given the same input and parameters, they will generally produce highly similar outputs, ensuring uniform stylistic adjustments in large volumes of text. Human editing: While professional editors strive for consistency, their work, being a human endeavour, will inherently have subtle variations influenced by their subjective interpretation, mood, and evolving understanding of the text. Depth of comprehension AI humanisers: Are fundamentally limited by their algorithms. They excel at pattern recognition and statistical likelihood but lack genuine comprehension, important thinking, or the ability to question the underlying meaning of the content. Human editing: Engages with the text on a profound level, understanding context, detecting logical fallacies, challenging assumptions, and ensuring the content’s intellectual rigour and persuasive power. Strategic implementation: Choosing your refinement path Deciding between an AI language humaniser and human editing isn’t about declaring a universal winner, but rather understanding which tool best suits the specific context and goals of your content. When to opt for an AI language humaniser Initial draft polishing: For rapidly transforming raw AI-generated content into a more palatable first or second draft. High-volume, low-stakes content: Ideal for social media captions, short informational blurbs, or SEO content where speed and quantity outweigh extreme stylistic originality. Overcoming AI “tells”: When the primary goal is simply to remove the most obvious signs of AI generation and make text flow more naturally. Budgetary constraints: When financial resources are limited and a quick, affordable improvement is needed. When to engage human editorial expertise High-stakes content: For academic theses, grant proposals, important business communications, or creative writing where precision, originality, and impact are paramount. Developing a unique voice: When the content needs to embody a distinctive brand persona, a specific authorial style, or convey complex emotional resonance. Complex or sensitive topics: For subjects requiring deep contextual understanding, cultural sensitivity, ethical nuance, or the ability to challenge underlying assumptions. Strategic communication: When the written piece is a key component of a broader strategy, requiring an editor’s insight into messaging, persuasion, and audience psychology. The synergistic approach: A modern workflow In many contemporary content pipelines, the most effective strategy isn’t an either/or but a blend of both. The hybrid workflow uses the strengths of each. AI can efficiently generate an initial draft and perform a preliminary humanisation pass, saving significant time. The refined AI output then becomes the starting point for a human editor. The human editor can then focus their invaluable skills on the higher-order concerns: injecting genuine creativity, ensuring perfect alignment with the author’s voice, refining cultural nuances, and providing a deep contextual understanding. The synergistic approach maximises efficiency without sacrificing the irreplaceable human touch that truly elevates content. Conclusion The debate between AI language humanisers and human editing is not a zero-sum game. It highlights an evolving ecosystem of content creation. AI tools, with their speed and cost-effectiveness, are incredibly adept at removing the robotic stiffness from machine-generated text, making it more palatable for general consumption. They provide a valuable first pass for efficiency. However, the depth, intuition, cultural awareness, and unique voice that a skilled human editor brings remain unparalleled. For content that truly needs to inspire, persuade, or connect on a profound human level, the human element is indispensable. The most forward-thinking creators will embrace a hybrid model, using AI for efficiency in initial stages and reserving the invaluable human touch for final polish, strategic refinement, and ensuring genuine connection. Author: Kateryna Bykova, StudyAgent. Image source: Pexels The post How do AI ‘humanisers’ compare to human editing? appeared first on AI News. View the full article

For chief data and information officers, especially in tightly regulated sectors, data governance has been a major cause preventing enterprise adoption of AI models. The issue of data sovereignty – which concerns where company data is handled and kept – has held many back, forcing them to use complex private cloud solutions. Others have simply given up. OpenAI’s recent announcement to offer *** data residency shows that big AI model providers are changing their setups to match the strict data protection and rules that enterprise and public sector clients require. This directly deals with the top governance concern in the market and should speed up AI adoption. It will move AI past test projects and into important business functions. From public sector tests to full adoption The new *** data residency choice, starting October 24, will apply to OpenAI’s main business products: the API Platform, ChatGPT Enterprise, and ChatGPT Edu. This lets *** clients keep their enterprise data in the ***, assisting with AI governance and meeting local data protection rules. The *** Ministry of Justice (MoJ) is the first major client for this. They’ve signed a deal to give 2,500 civil servants access to ChatGPT Enterprise. This is a full deployment after the MoJ said a recent trial run showed people saved time on routine tasks including writing help, compliance and legal jobs, research, and document work. The deal backs the department’s AI Action Plan and aims to help workers be more productive and better serve the public. This specific use in a government legal section gives a reliable example for others – such as those in finance and healthcare – to measure how much they can gain from using AI on complicated, knowledge-based tasks. Adoption by the MoJ joins other AI tools already in action in Whitehall, like ‘Humphrey,’ an AI assistant that helps with admin, and ‘Consult,’ a tool that looks at public feedback in minutes instead of weeks. What it takes to implement and challenges This announcement points out two paths for OpenAI’s *** setup plans. The new data residency choice looks like the fix for enterprise AI data governance right now. It is separate from Stargate ***, the earlier AI project with NVIDIA and Nscale. That one is about building sovereign AI by delivering models on local computing for specific uses in the long run. For IT leaders, this complicates the AI platform market, which already has many choices. OpenAI’s move to offer data residency goes against what cloud providers offer. Before, companies wanting to use OpenAI models within a specific place were pointed towards platforms like Microsoft’s Azure AI, which mixes model access with data governance. Now, they must consider things more. Businesses can access the models straight from OpenAI – getting new features and *** residency – or keep using platforms like Azure AI, AWS Bedrock, or Google Vertex AI. These may connect better with current data and company applications. This choice will be compared against platforms like IBM watsonx or AI in business software like SAP Joule, which focus on data privacy and fitting into current workflows. Sam Altman, CEO of OpenAI, said: “The number of people using our products in the *** has increased fourfold in the past year. It’s exciting to see them using AI to save time, increase productivity, and get more done. “Civil servants are using ChatGPT to improve public services and established firms are reimagining operations. We’re proud to continue supporting the *** and the Government’s AI plan.” *** Deputy Prime Minister David Lammy added: “Our partnership with OpenAI places Britain firmly in the driving seat of the global tech revolution—leading the world in innovation and using technology to deliver fairness and opportunity for every corner of the United Kingdom.” Key points for enterprise leaders navigating AI data governance OpenAI’s move from a US-focused setup to offering local data options responds to what companies and governments want. For business leaders, this change calls for a review. Look again at governance issues: CISOs and Data Protection Officers should check any risk reviews that blocked OpenAI tools because of data residency. This change might now permit AI projects. Consider government uses: The MoJ’s good test case offers a strong reason to use this change. CIOs and COOs in other fields can point to a government using this tech for work like document analysis, making a business case for investing. Think about total costs: CTOs should now compare the total cost of working directly with OpenAI against using a cloud platform. This should include API prices and the cost of integration, security, and meeting rules. Get ready for sovereign AI: This is part of a ******* trend. The Stargate *** project and the government’s MOU show that sovereign AI is a long-term goal. Business tech planners should get ready for a mix of AI where models and data are handled in different places, with some running locally for rules, speed, or security. Enterprise leaders can now more easily use the AI platform directly, since a big data governance issue has been solved. The question is no longer whether we can use AI tools securely, but how to include, manage, and grow them to get real business results. See also: How AI adoption is moving IT operations from reactive to proactive 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 Expo, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post OpenAI data residency advances enterprise AI governance appeared first on AI News. View the full article

Reports are circulating this week that Meta is cutting approximately 600 positions from its AI division, a move that seems paradoxical given the company’s aggressive recruitment campaign over recent months. The contradiction raises important questions about Meta’s AI strategy and what it signals for the broader tech industry. For those following Meta AI job cuts, the timing is striking. Just months after the company went on a highly publicised hiring spree – offering compensation packages reportedly reaching up to hundreds of millions of dollars to lure top researchers from OpenAI, Google, and other competitors – Meta is now scaling back parts of its AI workforce. The numbers behind Meta AI job cuts The cuts will affect Meta’sFAIR AI research, product-related AI, and AI infrastructure units in the company’s Superintelligence Labs, which employs several thousand people, according to a report byAxios. However, the newly-formed TBD Lab unit will be spared from cuts. Following the layoffs, Meta’s Superintelligence Labs’ workforce now sits at just under 3,000, CNBC stated. The company has offered affected employees 16 weeks of severance plus two weeks for every completed year of service, and is encouraging them to apply for other positions in Meta. Why is Meta making these cuts? According to Axios, an internal memo from Meta Chief AI Officer Alexandr Wang indicated that the restructuring aims to address what the company concluded was an overly bureaucratic structure. “By reducing the size of our team, fewer conversations will be required to make a decision, and each person will be more load-bearing and have more scope and impact,” Wang wrote. The backstory reveals deeper concerns. CEO Mark Zuckerberg grew concerned several months ago that the company’s existing AI efforts weren’t leading to needed breakthroughs or improved performance. The dissatisfaction reportedly stemmed from the lukewarm response to Meta’s Llama 4 models released in April. The expensive hiring spree To understand the current Meta AI job cuts, we need to examine what came before. In June 2025, Meta made a US$14.3 billion investment in Scale AI and brought on the startup’s CEO, Alexandr Wang, as Meta’s first-ever Chief AI Officer. The company then embarked on an aggressive talent acquisition campaign. Meta hired multiple researchers from OpenAI, including Shengjia Zhao, Jiahui Yu, Shuchao ***, and Hongyu Ren. Meta reportedly also poached more than 50 researchers from competitors, with OpenAI CEO Sam Altman claiming Meta was offering “US$100 million signing bonuses.” Zuckerberg stated he was focused on “building the most elite and talent-dense team in the industry” for Meta’s new Superintelligence Labs. The company also hired prominent executives, including former GitHub CEO Nat Friedman and Safe Superintelligence co-founder Daniel Gross. Then, unexpectedly, Meta paused hiring for its AI division in August 2025, just weeks after the massive recruitment push. The new guard vs. the old guard What makes these Meta AI job cuts particularly revealing is who’s being affected – and who isn’t. The cuts did not impact employees in TBD Labs, which includes many of the top-tier AI hires brought into Meta this summer, CNBC stated. In Meta, the AI unit was considered bloated, with teams like FAIR and product-oriented groups often vying for computing resources. The restructuring appears to be a calculated bet on the new talent over legacy teams. Timing that raises eyebrows The timing of these Meta AI job cuts is particularly noteworthy. Just a day before the layoffs were announced, Meta secured a US$27 billion financing deal with Blue Owl Capital to fund the Hyperion data centre in Louisiana. The juxtaposition is stark: Meta is simultaneously cutting AI personnel while investing tens of billions in AI infrastructure. This suggests the company isn’t pulling back from AI – it’s redirecting resources toward specific initiatives it deems more promising. What this means for the AI industry The Meta AI job cuts may signal a broader shift in the tech industry’s approach to AI talent. This raises questions in Silicon Valley about whether AI layoffs are beginning to surface just as the hype cycle peaks. After months of frenzied hiring and astronomical compensation packages, Meta’s restructuring suggests that simply accumulating AI talent isn’t enough. The company appears to be learning that organisational structure, decision-making speed, and team coherence matter as much as individual brilliance. Tech analyst Dan Ives of Wedbush Securities described Meta as being in “digestion mode” after a spending spree, while Daniel Newman, CEO of Futurum Group, called the hiring freeze “a natural resting point for Meta,” as per CNBC‘s report. The ******* picture Despite the layoffs, Meta insists its commitment to AI remains unwavering. The company continues to actively recruit and hire for its TBD Lab unit, and Zuckerberg has said Meta’s AI initiatives will result in a 2026 year-over-year expense growth rate above 2025’s growth. What we’re witnessing isn’t a retreat from AI but a strategic realignment. Meta is consolidating its AI efforts around a smaller, more agile core team led by Wang and populated by the expensive talent acquisitions from earlier this year. The company is betting that this leaner structure will deliver the breakthroughs that eluded its larger, more established teams. The bottom line The contradiction of Meta AI job cuts alongside continued hiring isn’t contradictory at all – it’s a deliberate strategy. Meta is cutting the old to make room for the new, streamlining bureaucracy, and betting that its expensive new hires will succeed where legacy teams struggled. Whether this gamble pays off remains to be seen. The company is creating a startup-in-a-giant, protecting its prized recruits yet trimming the organisational ****. As Wang noted in his memo, “This is a talented group of individuals, and we need their skills in other parts of the company.” Whether Meta can successfully redeploy this talent internally – or whether they’ll join competitors – will be another chapter in the ongoing AI talent wars. Meta’s approach reflects a broader truth about the AI industry: throwing money and people at the problem isn’t enough. Success requires the right structure, the right strategy, and increasingly, the courage to make difficult decisions about what – and who – to prioritise. See also: Zuckerberg outlines Meta’s AI vision for ‘personal superintelligence’ 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 Expo, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Meta hires and fires AI workers: Behind the contradiction appeared first on AI News. View the full article

Security experts at JFrog have found a ‘prompt **********’ threat that exploits weak spots in how AI systems talk to each other using MCP (Model Context Protocol). Business leaders want to make AI more helpful by directly using company data and tools. But, hooking AI up like this also opens up new security risks, not in the AI itself, but in how it’s all connected. This means CIOs and CISOs need to think about a new problem: keeping the data stream that feeds AI safe, just like they protect the AI itself. Why AI attacks targeting protocols like MCP are so dangerous AI models – no matter if they’re on Google, Amazon, or running on local devices – have a basic problem: they don’t know what’s happening right now. They only know what they were trained on. They don’t know what code a programmer is working on or what’s in a file on a computer. The boffins at Anthropic created the MCP to fix this. MCP is a way for AI to connect to the real world, letting it safely use local data and online services. It’s what lets an assistant like Claude understand what this means when you point to a piece of code and ask it to rework this. However, JFrog’s research shows that a certain way of using MCP has a prompt ********** weakness that can turn this dream AI tool into a nightmare security problem. Imagine that a programmer asks an AI assistant to recommend a standard Python tool for working with images. The AI should suggest Pillow, which is a good and popular choice. But, because of a flaw (CVE-2025-6515) in the oatpp-mcp system, someone could sneak into the user’s session. They could send their own fake request and the server would treat it like it came from the real user. So, the programmer gets a bad suggestion from the AI assistant recommending a fake tool called theBestImageProcessingPackage. This is a serious attack on the software supply chain. Someone could use this prompt ********** to inject bad code, steal data, or run commands, all while looking like a helpful part of the programmer’s toolkit. How this MCP prompt ********** attack works This prompt ********** attack messes with the way the system communicates using MCP, rather than the security of the AI itself. The specific weakness was found in the Oat++ C++ system’s MCP setup, which connects programs to the MCP standard. The issue is in how the system handles connections using Server-Sent Events (SSE). When a real user connects, the server gives them a session ID. However, the flawed function uses the computer’s memory address of the session as the session ID. This goes against the protocol’s rule that session IDs should be unique and cryptographically secure. This is a bad design because computers often reuse memory addresses to save resources. An attacker can take advantage of this by quickly creating and closing lots of sessions to record these predictable session IDs. Later, when a real user connects, they might get one of these recycled IDs that the attacker already has. Once the attacker has a valid session ID, they can send their own requests to the server. The server can’t tell the difference between the attacker and the real user, so it sends the malicious responses back to the real user’s connection. Even if some programs only accept certain responses, attackers can often get around this by sending lots of messages with common event numbers until one is accepted. This lets the attacker mess up the model’s behaviour without changing the AI model itself. Any company using oatpp-mcp with HTTP SSE enabled on a network that an attacker can access is at risk. What should AI security leaders do? The discovery of this MCP prompt ********** attack is a serious warning for all tech leaders, especially CISOs and CTOs, who are building or using AI assistants. As AI becomes more and more a part of our workflows through protocols like MCP, it also gains new risks. Keeping the area around the AI safe is now a top priority. Even though this specific CVE affects one system, the idea of prompt ********** is a general one. To protect against this and similar attacks, leaders need to set new rules for their AI systems. First, make sure all AI services use secure session management. Development teams need to make sure servers create session IDs using strong, random generators. This should be a must-have on any security checklist for AI programs. Using predictable identifiers like memory addresses is not okay. Second, strengthen the defenses on the user side. Client programs should be designed to reject any event that doesn’t match the expected IDs and types. Simple, incrementing event IDs are at risk of spraying attacks and need to be replaced with unpredictable identifiers that don’t collide. Finally, use zero-trust principles for AI protocols. Security teams need to check the entire AI setup, from the basic model to the protocols and middleware that connect it to data. These channels need strong session separation and expiration, like the session management used in web applications. This MCP prompt ********** attack is a perfect example of how a known web application problem, session **********, is showing up in a new and dangerous way in AI. Securing these new AI tools means applying these strong security basics to stop attacks at the protocol level. See also: How AI adoption is moving IT operations from reactive to proactive 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 Expo, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post MCP prompt **********: Examining the major AI security threat appeared first on AI News. View the full article

Link strategies are important in improving search engine optimisation (SEO) and online visibility. Artificial intelligence is changing these strategies, making them more accurate and efficient. This article examines AI’s impact on link strategies, highlighting current tools and future trends. Businesses are always looking for new ways to improve their online presence. Link strategies are the key to boosting SEO rankings, ensuring websites get the visibility they need. With AI integration, businesses can achieve better results by using data-driven insights and automation. This guide explores Bazoom’s backlink building service, showing how it can change traditional methods and help companies prepare for the future. How AI is reshaping link strategies Traditional link strategies often relied on manual processes, like contacting potential partners and tracking backlinks. These methods, while useful, were labour-intensive and prone to error. AI addresses these issues by automating many aspects of link building. Machine learning algorithms allow AI to quickly analyse large data sets, identifying the most relevant linking opportunities for your business. By using AI technologies, businesses can improve the accuracy of their backlink profiles. AI tools can predict which links will most impact search rankings based on historical data and current trends. The predictive ability helps businesses allocate resources more effectively and focus on high-value opportunities. Besides accuracy, AI saves time by automating repetitive tasks, allowing professionals to focus on strategy development. AI-enhanced link strategies offer benefits beyond efficiency and accuracy. The technologies provide insights into competitive landscapes, helping businesses understand their position relative to competitors. By gaining a comprehensive market view, companies can make informed decisions about their link building efforts and adapt quickly to changes in search engine algorithms. AI tools that enhance link strategies Today, various AI tools significantly improve link building processes. Tools use machine learning to identify data patterns that may not be immediately apparent to humans. Features like automated outreach, relationship management, and real-time analytics streamline the process from start to finish. AI tools offer more than automation. They provide actionable insights that help optimise link strategies for maximum impact. By analysing competitor links and industry trends, the tools offer strategic recommendations tailored to specific business needs. Integrating such technology enables tracking and reporting, providing visibility into each campaign’s effectiveness. For businesses aiming to stay competitive in SEO, implementing advanced tools is essential. They simplify complex processes and ensure every decision is backed by solid data analysis. The ability to forecast potential outcomes allows companies to adjust their strategies proactively rather than reactively. Practical uses of AI in link strategies AI’s practical applications in link strategies are diverse. For example, businesses can automate their backlink initiatives by using tools that facilitate efficient outreach and engagement with potential partners. The automation ensures consistent and targeted communication without overwhelming resources. AI’s impact on data analysis is significant; it transforms raw data into valuable insights that drive decision-making in SEO campaigns. By identifying patterns and predicting search engine trends, AI helps businesses optimise their link building efforts effectively. The capability ensures companies remain competitive in a constantly changing digital landscape. Moreover, AI enables businesses to predict future industry trends accurately. By forecasting shifts in consumer behaviour or search engine algorithm changes, companies can adjust their strategies accordingly and maintain a competitive edge. The proactive approach reduces the risk associated with sudden changes in SEO dynamics. Future trends in AI and link building practices Several emerging trends suggest further integration between AI technologies and digital marketing efforts. One notable tendency is the increased use of natural language processing (NLP) in SEO tools, allowing for better understanding of contextually relevant content creation. The potential for AI integration extends beyond improving existing systems; it offers opportunities for creating new marketing paradigms through collaborative approaches in various platforms like social media channels or content management systems. As these technologies continue to advance, expect more innovation in the future. While predicting what’s around the corner can be challenging, one thing is clear: embracing technology advances will be crucial for success. Adopting cutting-edge solutions now ensures readiness for whatever comes next, and provides a foundation for sustainable growth in the long term. 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 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 AI is changing how we build links for SEO appeared first on AI News. View the full article

CIOs want to fix IT problems faster without expanding headcount, and many see AI adoption as the solution for their operations. For ages, they’ve used things like automation and self-help portals to handle this, so their teams can solve issues quickly. Now, AI is getting involved, and lots of companies are trying to use it for their IT support. There’s been a ton of buzz, but leaders want proof that it actually works. SolarWinds looked into how these new tools are doing. The company analysed a bunch of info from over 2,000 IT systems and 60,000 pieces of data collected over a year, from August 2024 to July 2025. For their study, SolarWinds checked out AI stuff that’s supposed to make things easier, like automatically suggesting answers to tickets, finding helpful articles, and making summaries of problems. The results give a good idea of how much more efficient companies can get. How much more efficient does AI make IT operations? The main thing the report found is that it takes way less time to fix an IT issue after a company starts using AI. Before AI, it took about 27.42 hours to solve a problem. After, it dropped to 22.55 hours. That’s about 17.8 percent faster, saving about 4.87 hours per problem. This lets IT teams spend more time on tricky stuff instead of getting bogged down with everyday issues. This can save companies a bunch of money. The report talks about a medium-sized IT team handling 5,000 problems a year. By saving about 4.87 hours on each one, they’d get back 24,350 hours of work each year. If you figure a help desk person costs $28 an hour, that’s like saving over $680,000. But it’s not just about saving money. The report says IT can use that time to work on important projects and fix problems before they happen. This helps IT go from just fixing things to actually helping the company do better. The report also shows that there’s a big difference between companies that use AI and those that don’t. Companies using AI fix tickets in about 22.55 hours, while others take about 32.46 hours. That’s about a 30.5 percent difference which means almost 10 hours saved per problem for those using AI. However, it’s not just about the AI The report makes it clear that AI isn’t a magic fix for IT operations. It only works if you have good processes and are ready to make broader, company-wide changes. The best example is a group called the ‘Top 10 AI Adopters’. These ten companies stood out because they cut their resolution times the most. They went from about 51 hours to 23 hours, which is more than half the time saved. Their secret wasn’t special software, but how they used it. These companies all had one thing in common: they didn’t just try out AI as a side project. They made it a part of their daily work to help fix problems. The report basically says that AI works best when you also make changes to your processes and are willing to improve things. The report also says that it helps to have a culture that’s already into things like self-service portals and automation. These teams are already trying to make their support desks strong and ready for anything. AI just works better when you have these things in place. What to do if you’re in charge The SolarWinds report shows that AI for IT support operations isn’t just a possibility, it can really work. Cutting resolution times by almost 18 percent is a big deal for leaders. Here’s what they should do: See where you’re at: Before you spend money, figure out how long it takes to fix things now. The report says the average for companies not using AI is about 32.46 hours. Knowing your own number helps you decide if it’s worth it. Make it part of the job: The top users show that it’s better to use AI every day than to just try it out on the side. This means making changes to how you work and focusing on making things better. AI is a tool, not a miracle: AI can really help speed things up, especially if you already have good IT practices. Check your knowledge base and automation rules. AI works best when you have clear processes in place. Figure out how much you can save: The report gives a simple way to show your team how much time and money AI can save. Just multiply the number of incidents you have each year by the average saving of 4.87 hours. This gives you a clear idea of how much more efficient you can be. The difference between companies using AI and those who aren’t is growing. Leaders need to set up their operations so they can use AI to help IT become a real partner in the company’s success. See also: China’s generative AI user base doubles to 515 million in six months 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 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 AI adoption is moving IT operations from reactive to proactive appeared first on AI News. View the full article

A few years ago, the business technology world’s favourite buzzword was ‘Big Data’ – a reference to organisations’ mass collection of information that could be used to suggest previously unexplored ways of operating, and float ideas about what strategies they may best pursue. What’s becoming increasingly apparent is that the problems companies faced in using Big Data to their advantage still remain, and it’s a new technology – AI – that’s making those problems rise once again to the surface. Without tackling the problems that beset Big Data, AI implementations will continue to fail. So what are the issues stopping AI deliver on its promises? The vast majority of problems stem from the data resources themselves. To understand the issue, consider the following sources of information used in a very average working day. In a small-to-medium sized business: Spreadsheets, stored on users’ laptops, in Google Sheets, Office 365 cloud. The customer relationship manager (CRM) platform. Email exchanges between colleagues, customers, suppliers. Word documents, PDFs, web forms. Messaging apps. In an enterprise business: All of the above, plus, Enterprise resource planning (ERP) systems. Real-time data feeds. Data lakes. Disparate databases behind multiple point-products. It’s worth noting that the simple list above isn’t comprehensive, and nor is it intended to be. What it demonstrates is that in just five lines, there are around a dozen places where information can be found. What Big Data needed (perhaps still needs) and what AI projects also rest on, is somehow bringing all those elements together in such a way that a computer algorithm can make sense of it. Marketing behemoth Gartner’s hype cycle for artificial intelligence, 2024, placed AI-Ready Data on the upward curve of the hype cycle, estimating it would be 2-5 years before it reached the ‘plateau of productivity’. Given that AI systems mine and extract data, most organisations – save those of the very largest size – don’t have the foundations on which to build, and may not have AI assistance in the endeavour for another 1-4 years. The underlying problem for AI implementation is the same as dogged Big Data innovations as they, in the past, made their way through the hype cycle – from innovation trigger, peak of inflated expectations, trough of disillusionment, slope of enlightenment, to plateau of productivity – data comes in many forms; it can be inconsistent; perhaps it adheres to different standards; it may be inaccurate or biased; it could be highly sensitive information, or old and therefore irrelevant. Transforming data so it’s AI-ready remains a process that’s as relevant today (perhaps more so) than it’s ever been. Those companies wanting to get a jump start could experiment with the many data treatment platforms currently available, and as is becoming the common advice, might begin with discrete projects as test-beds to assess the effectiveness of emerging technologies. The advantage of the latest data preparation and assembly systems is that they are designed to prepare an organisation’s information resources in ways that are designed for the data to be used by AI value-creation platforms. They can offer, for example, carefully-coded guardrails that will help ensure data compliance, and protect users from accessing biased or commercially-sensitive information. But the challenge of producing coherent, safe, and well-formulated data resources remains an ongoing issue. As organisations gain more data in their everyday operations, compiling up-to-date data resources on which to draw is a constant process. Where big data could be considered a static asset, data for AI ingestion has to be prepared and treated in as close to real-time as possible. The situation therefore remains a three-way balance between opportunity, risk, and cost. Never before has the choice of vendor or platform been so crucial to the modern business. (Source: “Inside the business school” by Darien and Neil is licensed under CC BY-NC 2.0.) 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 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 Businesses still face the AI data challenge appeared first on AI News. View the full article

For CFOs and CIOs under pressure to modernise finance operations, automation, as seen in several generations of RPA (robotic process automation), isn’t enough. It’s apparent that transparency and explainability matter just as much. Accounting firms and finance functions inside organisations are now turning to AI systems that reason, not just compute. One of the most ambitious examples is Basis, a US-based start-up founded just two years ago that builds AI agents designed to automate structured accounting work, and keep human oversight firmly in the loop. Such systems signal a shift in enterprise automation. Instead of replacing people, AI agents extend human expertise and combine the precision of AI models with the oversight finance professionals need for compliance and client trust. Business impact: efficiency with accountability Basis develops AI agents that handle routine finance tasks such as reconciliations, journal entries, and financial summaries. The platform is built on OpenAI’s GPT-4.1 and GPT-5 models, models that give the facility to operators to examine each decision step taken autonomously. Accounting firms using Basis report up to 30% time savings and an ensuing higher capacity for advisory work. That’s the kind of value creation traditional automation cannot deliver as quickly or at similar cost to the business. Unlike many automation tools that operate as ****** boxes, Basis emphasises review-able reasoning. Every recommendation includes an account of the data used and the logic behind it. Visibility means accountants can validate each outcome and remain responsible for results, a feature that’s always important in financial operations, and especially in highly-regulated industries. Implementation and challenges: building systems that learn Agentic AI can treat accounting as a network of workflows, not isolated tasks. A supervising AI agent, powered by GPT-5 in the case of Basis’s platform, manages the entirety of processes. It can delegate specific tasks sub-agents running on different models, with the choice of AI model depending on the job’s complexity and the type of data that’s to be processed. For example, for quick queries or clarifications, Basis uses GPT-4.1 for its speed, while for complex classifications or month-end close, GPT-5 provides better reasoning and context handling. Company benchmarks each of its models against real-world accounting workflows to decide when it’s safe to let agents handle more responsibility. Finance professionals can always see what the system has done, why it made specific choices, and how confident it is in its recommendations. This malleable architecture lets firms scale AI and help ensure accuracy as levels of automation increase. The process mirrors the hybrid human–AI collaboration now emerging as the norm in sectors like legal services and risk management. Lessons for other sectors What makes Basis and financial multi-agentic AI relevant beyond accounting is the model-orchestration approach, routing tasks to the most appropriate AI model based on its performance and latency. The format could inform similar deployments in procurement, HR, or compliance operations; anywhere, in fact, where large volumes of structured decisions need transparency and – to use a terrible pun – accountability. Basis’s collaboration with OpenAI shows how AI reasoning engines in secure data environments can be effective. The goal isn’t pure speed, but automation that increases trust in the operator, and in the models themselves. These are systems that evolve without humans losing control of the outcomes. Conclusion AI in accounting isn’t limited to automating entries, it’s turning more towards building systems that think like accountants, not machines. For enterprise leaders, Basis’s model shows a way toward automation that improves over time. Each improvement in model makes teams faster and smarter without surrendering control to the automation process. (Image source: “Accounting charts” by World Bank Photo Collection is licensed under CC BY-NC-ND 2.0.) 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 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 How accounting firms are using AI agents to reclaim time and trust appeared first on AI News. View the full article

The AI adoption in China has reached unprecedented levels, with the country’s generative artificial intelligence user base doubling to 515 million in just six months, according to a report released by the China Internet Network Information Centre (CNNIC). This dramatic expansion represents an adoption rate of 36.5% in the first half of 2025, positioning China as a formidable force in the global AI landscape. The CNNIC report, published on Saturday, reveals that China’s AI adoption more than doubled between the end of December 2024 and June 2025. This growth trajectory underscores the rapid integration of AI technologies into ******** society, driven by what the report describes as “advanced infrastructure and state encouragement.” Demographic profile of users The survey of 30,000 respondents across 31 provinces reveals distinct patterns in who is embracing these technologies. Young and middle-aged professionals dominate the user base, with those under 40 years old accounting for 74.6% of all users. Education levels also play a significant role, as individuals with degrees from higher education institutions represent 37.5% of the total user base. This demographic concentration suggests that in China, the AI adoption is currently strongest among digitally native, educated populations who are positioned to leverage these tools in professional and personal contexts. Domestic models dominate market preferences Perhaps most notably, the report found that over 90% of respondents indicated their first choice was a domestic AI model. This preference for homegrown solutions reflects both Beijing’s strategic emphasis on technological self-reliance and the practical reality that leading American models from OpenAI and Google DeepMind are officially blocked on the mainland. ******** AI platforms have filled this void effectively. Models such as DeepSeek, Alibaba Cloud’s Qwen, and ByteDance’s Doubao have “exploded in popularity among ******** users,” according to a South China Morning Post report. Alibaba Cloud serves as the AI and cloud computing unit of Alibaba Group Holding. Global context and comparative analysis Despite restrictions on access to US-based AI models, China’s market has already established itself as the world’s largest. A Microsoft study published in September found that China had an estimated AI user base exceeding 195 million as of June 2024. The company’s research utilised a metric that quantifies China’s AI adoption as the share of the working-age population actively using AI tools. The Microsoft study identified a crucial inflexion point: the launch of DeepSeek’s R1 model in January 2025 led to AI adoption in China more than doubling to 20% in the subsequent six months. By comparison, the United States maintained a relatively steady adoption rate of approximately 25% over the past year. Zhang Xiao, deputy director of the CNNIC, told state news agency Xinhua that further innovations are anticipated in “open source AI, embodied intelligence, AI agents and AI governance.” Innovation and patent leadership Beyond user adoption metrics, China has established significant leadership in AI innovation as measured by patent applications. The CNNIC report noted that China had filed 1.576 million AI-related patent applications as of April 2025, representing 38.58% of the global total—the highest share of any country worldwide. This patent activity suggests that the AI adoption in China extends beyond consumer usage to encompass substantial research and development efforts across the technology stack. Strategic policy framework The rapid expansion in China’s AI adoption aligns with Beijing’s “AI Plus” initiative, which calls for widespread diffusion of AI technologies throughout society and the economy. The government has emphasised the importance of self-reliance across the entire AI ecosystem, from foundational models to computing hardware. This policy framework has created an environment where domestic technology companies are incentivised to develop competitive AI solutions while users are encouraged to integrate these tools into their daily workflows. Implications for the global AI landscape The doubling of China’s generative AI user base to 515 million users in six months represents more than a statistical milestone. It signals the emergence of a parallel AI ecosystem that operates largely independently of Western platforms yet serves a massive user population. As China’s AI adoption continues to accelerate, the global technology landscape may increasingly feature two distinct spheres of influence—one centred on American models and platforms, another on ******** alternatives. This bifurcation could have profound implications for how AI technologies evolve, how innovation diffuses across borders, and how global standards for AI governance take shape. The coming months will reveal whether this growth trajectory can be sustained and how effectively ******** AI platforms can continue to meet user needs in an increasingly competitive and rapidly evolving market. See also: Will the budget China AI chip from Nvidia survive Huawei’s growth? 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 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 generative AI user base doubles to 515 million in six months appeared first on AI News. View the full article

Arm has announced that it’s providing its most powerful edge AI platform, Armv9, to startups via its Flexible Access programme. The “Flexible Access” model is essentially a ‘try before you buy’ for chip designers. It gives companies upfront, low-cost, or no-cost (for qualifying startups) access to a wide range of Arm technology, tools, and resources. They can experiment and iterate designs freely while only paying license fees for the technology they use in final designs. This approach has already been a “catalyst for innovation,” according to Arm. The model helped to create around 400 successful chip designs (or “tape-outs”) over the last five years. You’ve likely heard of some of the companies already using it, like Raspberry Pi, Hailo, and SiMa.ai. The Armv9 edge AI platform pairs the super-efficient Arm Cortex-A320 processor with the Arm Ethos-U85 NPU, which is the bit that handles the heavy AI lifting. This duo is capable of running AI models with over one billion parameters right on the device itself, no cloud connection needed. This is the tech that will power the next-gen edge AI applications such as smart cameras that don’t just record but understand what they’re seeing, smart home gadgets that learn your habits, and robots you can interact with using vision, voice, and gesture. Paul Williamson, who runs the IoT business at Arm, believes the next wave of AI innovation will happen “at the edge – in the devices, interfaces, and systems that bring intelligence closer to where data is created.” A huge benefit here is privacy and security. By processing everything locally, machines can “perceive and respond like humans, while keeping inference and data processing securely on-device”. Your personal data doesn’t have to be sent off to a server just to figure out what you said. The Armv9 platform also bakes in security features like Pointer Authentication Code (PAC) and Memory Tagging Extension (MTE) to keep that on-device data safe. Research from VDC predicts that, by 2028, AI will be the “dominant technology used across IoT projects”. Arm’s technology is “already at the center of this transformation,” and this move just solidifies its position. For all the developers keen to get started with edge AI, the Arm Cortex-A320 will be available through the programme in November 2025, with the Ethos-U85 AI processor following in early 2026. See also: NVIDIA GPUs to power Oracle’s next-gen enterprise AI services 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 Expo, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Arm provides edge AI platform to startups via flexible access appeared first on AI News. View the full article

Google has announced DeepSomatic, an AI tool that can identify *******-related mutations in tumour genetic sequences more accurately. ******* starts when the controls governing cell division malfunction. Finding the specific genetic mutations driving a tumour’s growth is essential for creating effective treatment plans. Doctors now regularly sequence tumour cell genomes from biopsies to inform treatments that can target how a particular ******* grows and spreads. Published in Nature Biotechnology, this work presents a tool that uses convolutional neural networks to identify genetic variants in tumour cells with greater accuracy than current methods. Google has made both DeepSomatic and the high-quality training dataset created for it openly available. The challenge of somatic variants ******* genetics is complex. While genome sequencing finds genetic ******* variations, distinguishing real variants from sequencing errors is difficult and where an AI tool would provide welcome assistance. Most cancers are driven by ‘somatic’ variants acquired after birth rather than inherited ‘germline’ variants from parents. Somatic mutations happen when environmental factors like UV light damage DNA, or when random errors occur during DNA replication. When these variants alter normal cell behaviour, they can cause uncontrolled replication, driving ******* development and progression. Identifying somatic variants is harder than finding inherited ones because they can exist at low frequencies within tumour cells, sometimes at rates lower than the sequencing error rate itself. How DeepSomatic works In clinical settings, scientists sequence both tumour cells from a biopsy and normal cells from the patient. DeepSomatic spots the differences, identifying variations in tumour cells that aren’t inherited. These variations reveal what’s fuelling the tumour’s growth. The model converts raw genetic sequencing data from both tumour and normal samples into images representing various data points, including the sequencing data and its alignment along the chromosome. A convolutional neural network analyses these images to differentiate between the standard reference genome, the individual’s normal inherited variants, and *******-causing somatic variants while filtering out sequencing errors. The output is a list of *******-related mutations. DeepSomatic can also work in ‘tumour-only’ mode when normal cell samples are unavailable, which happens frequently with blood cancers like leukaemia. This makes the tool applicable across many research and clinical scenarios. Training a more precise AI ******* research tool Training an accurate AI model requires high-quality data. For its AI tool, Google and its partners at the UC Santa Cruz Genomics Institute and the National ******* Institute created a benchmark dataset called CASTLE. They sequenced tumour and normal cells from four breast ******* samples and two lung ******* samples. These samples were analysed using three leading sequencing platforms to create a single, accurate reference dataset by combining the outputs and removing platform-specific errors. The data shows how even the same ******* type can have vastly different mutational signatures, information that can help predict patient response to specific treatments. DeepSomatic models performed better than other established methods across all three major sequencing platforms. The tool excelled at identifying complex mutations called insertions and deletions, or ‘Indels’. For these variants, DeepSomatic achieved a 90% F1-score on Illumina sequencing data, compared to 80% for the next-best method. The improvement was more dramatic on Pacific Biosciences data, where DeepSomatic scored over 80% while the next-best tool scored less than 50%. The AI performed well when analysing challenging samples. Testing included a breast ******* sample preserved with formalin-fixed-paraffin-embedded (FFPE), a common method that can introduce DNA damage and complicate analysis. It was also tested on data from whole exome sequencing (WES), a more affordable method that sequences only the 1% of the genome coding for proteins. In both scenarios, DeepSomatic outperformed other tools, suggesting its utility for analysing lower-quality or historical samples. An AI tool for all cancers The AI tool has shown it can apply its learning to new ******* types it wasn’t trained on. When used to analyse a glioblastoma sample, an aggressive brain *******, it successfully pinpointed the few variants known to drive the disease. In a partnership with Children’s Mercy in Kansas City, it analysed eight samples of paediatric leukaemia and found the previously known variants while identifying 10 new ones, despite working with tumour-only samples. Google hopes research labs and clinicians will adopt this tool to better understand individual tumours. By detecting known ******* variants, it could help guide choices for existing treatments. By identifying new ones, it could lead to new therapies. The goal is to advance precision medicine and deliver more effective treatments to patients. See also: MHRA fast-tracks next wave of AI tools for patient care 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 Expo, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Google AI tool pinpoints genetic drivers of ******* appeared first on AI News. View the full article

The South Korean government spent 1.2 trillion won ($850m) on developing AI textbooks for schools, but the national programme has been rolled back after just four months, amid allegations of inaccurate texts, concerns about privacy, and increased workloads on staff and pupils. Writing in Rest Of World, journalist Junhyup Kwon quotes a student as saying, “All our classes were delayed because of technical problems with the textbooks. […] I found it hard to stay focused and keep on track. The textbooks didn’t provide lessons tailored to my level.” Kim Jong-hee, chief digital officer of *****-A Publishing, one of the textbook developers, spoke of the advantages of AI books: “Using digital devices [students] are familiar with keeps them more focused, awake, and more willing to participate. The textbooks provide more personalised support for students struggling with lessons.” The Korean government originally commissioned publishers to produce the AI textbooks, who in turn spent around $567m to develop the online, digital texts. The use of AI textbooks was made mandatory in the country from the beginning of the school year in March, but has since been classed as ‘optional’ after just one semester. The number of schools using the AI textbooks has halved in that time. Speaking in the National Assembly in January this year, legislator Kang Kyung-sook asked the Minister for Education, “Traditional print textbooks take 18 months to develop, nine months for review, and six months for preparation. But the AI textbooks took only 12 [months to develop], three [months for review], and three months [for preparation] […]. Why was it rushed? Since they target children, they require careful verification and careful procedures.” The failure of the AI textbook scheme has also been blamed on the politicisation of the issue, and a change of government as the programme was being rolled out. Technology programmes in schools since the widespread adoption of the internet are relatively common, have cost taxpayers considerably less, and lasted much longer – despite eventual failure or wholesale realignment. In South Africa’s Guateng Province in the early 2000s, the Online Schools Project was designed to equip schools with computer labs and internet connections, but was scrapped in 2013 at a cost of R1-billion rand ($57m), according to some reports. In 2019, Malaysia’s 1BestariNet – a cloud-based VLE (virtual learning environment) – was terminated after eight years amid investigations into alleged inconsistencies between internet speed claims and the reality experienced by many schools. The overall cost of that project was put in the billions of ringgit (one billion ringgit is around $235m). However, the speed of the failure of the South Korean AI textbooks project and its high cost, suggest the educational adoption of AI texts delivered digitally is pitted with difficulty. An academic study conducted by the Massachusetts Institute of Technology published earlier this year hinted that using AI in educational contexts lowers brain activity in the long-term, which suggests the technology may not be suitable for developing minds. (Image source: “Adorable sleeping students in the undergraduate library” by benchilada is licensed under CC BY-NC-SA 2.0.) 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 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 South Korea scraps AI textbook programme appeared first on AI News. View the full article

The surge of multi-billion-dollar investments in AI has sparked growing debate over whether the industry is heading for a bubble similar to the dot-com *****. Investors are watching closely for signs that enthusiasm might be fading or that the heavy spending on infrastructure and chips is failing to deliver expected returns. A recent survey by BofA Global Research found that 54% of fund managers believe AI stocks are already in bubble territory, while 38% disagree. Echoes of the dot-com era Despite the optimism surrounding AI, sceptics remain unconvinced of its real-world impact. Some even call it a bluff or a bubble waiting to burst. Speaking during Cisco’s recent Virtual Media Roundtable — AI Readiness Index 2025: Readiness Leads to Value, Ben Dawson, Senior Vice President and President for Asia Pacific, Japan, and Greater China (APJC), compared the current wave of AI hype to the early days of the internet. He said technological shifts of this scale often follow a familiar pattern — early excitement, heavy investment, and eventual market correction before long-term value takes hold. Dawson noted that while some AI projects or business models may not last, the overall transformation is real and lasting. He added that, much like the internet revolution, AI will permanently reshape business and society, and organisations that ignore it do so at their own risk. The role of governments and global policy Public policy is also shaping how the AI cycle unfolds — and how governments might cushion the risks of a potential AI bubble. As Harvard Business Review pointed out, in the US, government involvement has helped define past technology eras — often through incentives and early investments that encourage private innovation. The same pattern is now visible in AI. Both the Trump and Biden administrations have positioned AI as a matter of economic strength and national security, sending a clear message that speed matters. China has taken a state-led approach, directing capital toward local AI firms to reduce reliance on US technology. In Europe, efforts have focused more on regulation, though fears of overregulation have led to new programs — such as the AI Continent Action Plan and a €1 billion Apply AI fund — to boost adoption and competitiveness. Meanwhile, venture capital and sovereign wealth funds are investing heavily, even before widespread AI demand exists. These early bets assume that adoption will eventually justify the buildout. But if that demand slows, some investors could be left with stranded assets, much like the unused fibre networks that followed the dot-com bubble. For businesses, the challenge is different. Instead of financing the next infrastructure wave, they face the question of how to use AI to strengthen their operations. The companies that survived the dot-com downturn — such as Amazon — succeeded by aligning technology with real business value rather than market hype. Market warnings over a possible AI bubble The Bank of England recently warned that markets could suffer a sharp correction if confidence in AI falters, calling the potential impact on the ***’s financial system “material.” The warning reflects growing caution among policymakers about how quickly AI-related valuations have climbed. This concern is shared by some investors and economists who believe the rapid pace of AI spending may outstrip short-term returns. Others, however, argue that building AI infrastructure now is essential groundwork for future innovation. Building long-term AI infrastructure amid bubble fears When asked whether companies are worried about AI infrastructure costs and energy demand, Simon Miceli, Managing Director of Cloud and AI Infrastructure for APJC at Cisco, said he views the issue from the opposite angle. Rather than fearing overcapacity, he said what’s happening now is a large-scale buildout to support the industrialisation of AI. The question, he said, isn’t whether AI demand exists today, but whether the world is preparing fast enough for what’s coming. Miceli acknowledged that some correction in the AI market is likely, but he believes the long-term need for AI computing power justifies current investment levels. “There’s a race to develop AI and build the capability behind it,” he said, adding that demand will eventually meet supply as applications mature. Different shades of caution Across the industry, opinions vary on whether AI’s momentum represents hype or healthy growth. According to Reuters, at the Milken Institute Asia Summit 2025, Singapore’s GIC Chief Investment Officer Bryan Yeo said valuations in early-stage AI ventures appear inflated, with many startups commanding “huge multiples” despite modest revenues. He suggested that while some firms may justify their valuations, others are unlikely to deliver returns that match investor expectations. Jeff Bezos, Amazon’s founder, said that during periods of excitement like this, investors often struggle to separate good ideas from bad ones — though he also noted that innovation-driven bubbles often leave behind real progress once the market settles. At Goldman Sachs, economist Joseph Briggs argued that the current surge in AI infrastructure spending remains economically sustainable. He said the long-term case for AI investment is strong, but the ultimate winners are still uncertain given how quickly technology changes and how easily companies can switch providers. Meanwhile, ABB CEO Morten Wierod told Reuters that while he doesn’t see an AI bubble, supply chain and construction limits could slow the rollout of new data centres. IMF Chief Economist Pierre-Olivier Gourinchas added that even if there’s a downturn, it’s unlikely to cause a systemic financial crisis since AI investments aren’t debt-driven. OpenAI CEO Sam Altman also acknowledged market overexcitement, predicting that some investors will lose large sums while others will profit heavily — an outcome that mirrors past technology bubbles. Despite growing talk of an AI bubble, many investors remain committed to the sector. UBS equity strategists said that about 90% of investors who think the market is overheated are still holding AI-related assets, suggesting most believe the industry has not yet peaked. A cycle, not a collapse While concerns about an AI bubble are valid, most experts agree that the technology’s long-term impact is undeniable. As Cisco’s Ben Dawson put it, every major technological transition goes through a cycle of hype, correction, and consolidation — but what remains afterward reshapes industries for decades. For now, the question isn’t whether AI will endure, but how well businesses and investors can navigate the growing pains that come with every market bubble. (Photo by Growtika) See also: NVIDIA GPUs to power Oracle’s next-gen enterprise AI services 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 What if AI is the next dot-com bubble? appeared first on AI News. View the full article

Ant Group has entered the trillion-parameter AI model arena with Ling-1T, a newly open-sourced language model that the ******** fintech giant positions as a breakthrough in balancing computational efficiency with advanced reasoning capabilities. The October 9 announcement marks a significant milestone for the Alipay operator, which has been rapidly building out its artificial intelligence infrastructure across multiple model architectures. The trillion-parameter AI model demonstrates competitive performance on complex mathematical reasoning tasks, achieving 70.42% accuracy on the 2025 American Invitational Mathematics Examination (AIME) benchmark—a standard used to evaluate AI systems’ problem-solving abilities. According to Ant Group’s technical specifications, Ling-1T maintains this performance level while consuming an average of over 4,000 output tokens per problem, placing it alongside what the company describes as “best-in-class AI models” in terms of result quality. Dual-pronged approach to AI advancement The trillion-parameter AI model release coincides with Ant Group’s launch of dInfer, a specialised inference framework engineered for diffusion language models. This parallel release strategy reflects the company’s bet on multiple technological approaches rather than a single architectural paradigm. Diffusion language models represent a departure from the autoregressive systems that underpin widely used chatbots like ChatGPT. Unlike sequential text generation, diffusion models produce outputs in parallel—an approach already prevalent in image and video generation tools but less common in language processing. Ant Group’s performance metrics for dInfer suggest substantial efficiency gains. Testing on the company’s LLaDA-MoE diffusion model yielded 1,011 tokens per second on the HumanEval coding benchmark, versus 91 tokens per second for Nvidia’s Fast-dLLM framework and 294 for Alibaba’s Qwen-2.5-3B model running on vLLM infrastructure. “We believe that dInfer provides both a practical toolkit and a standardised platform to accelerate research and development in the rapidly growing field of dLLMs,” researchers at Ant Group noted in accompanying technical documentation. Ecosystem expansion beyond language models The Ling-1T trillion-parameter AI model sits within a broader family of AI systems that Ant Group has assembled over recent months. The company’s portfolio now spans three primary series: the Ling non-thinking models for standard language tasks, Ring thinking models designed for complex reasoning (including the previously released Ring-1T-preview), and Ming multimodal models capable of processing images, text, audio, and video. This diversified approach extends to an experimental model designated LLaDA-MoE, which employs Mixture-of-Experts (MoE) architecture—a technique that activates only relevant portions of a large model for specific tasks, theoretically improving efficiency. He Zhengyu, chief technology officer at Ant Group, articulated the company’s positioning around these releases. “At Ant Group, we believe Artificial General Intelligence (AGI) should be a public good—a shared milestone for humanity’s intelligent future,” He stated, adding that the open-source releases of both the trillion-parameter AI model and Ring-1T-preview represent steps toward “open and collaborative advancement.” Competitive dynamics in a constrained environment The timing and nature of Ant Group’s releases illuminate strategic calculations within China’s AI sector. With access to cutting-edge semiconductor technology limited by export restrictions, ******** technology firms have increasingly emphasised algorithmic innovation and software optimisation as competitive differentiators. ByteDance, parent company of TikTok, similarly introduced a diffusion language model called Seed Diffusion Preview in July, claiming five-fold speed improvements over comparable autoregressive architectures. These parallel efforts suggest industry-wide interest in alternative model paradigms that might offer efficiency advantages. However, the practical adoption trajectory for diffusion language models remains uncertain. Autoregressive systems continue dominating commercial deployments due to proven performance in natural language understanding and generation—the core requirements for customer-facing applications. Open-source strategy as market positioning By making the trillion-parameter AI model publicly available alongside the dInfer framework, Ant Group is pursuing a collaborative development model that contrasts with the closed approaches of some competitors. This strategy potentially accelerates innovation while positioning Ant’s technologies as foundational infrastructure for the broader AI community. The company is simultaneously developing AWorld, a framework intended to support continual learning in autonomous AI agents—systems designed to complete tasks independently on behalf of users. Whether these combined efforts can establish Ant Group as a significant force in global AI development depends partly on real-world validation of the performance claims and partly on adoption rates among developers seeking alternatives to established platforms. The trillion-parameter AI model’s open-source nature may facilitate this validation process while building a community of users invested in the technology’s success. For now, the releases demonstrate that major ******** technology firms view the current AI landscape as fluid enough to accommodate new entrants willing to innovate across multiple dimensions simultaneously. See also: Ant Group uses domestic chips to train AI models and cut costs 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 Expo, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Trillion-parameter AI model from Ant Group targets reasoning benchmarks with dual release strategy appeared first on AI News. View the full article

Healthcare regulator MHRA is fast-tracking new AI tools that promise to dramatically improve patient care. The wait for medical test results can stretch from days, to weeks, or even months. That wait ******* is often filled with worry and it always feels like an eternity. But what if that wait could be cut from weeks down to just a few minutes? In the next phase of its ‘AI Airlock’ programme, the Medicines and Healthcare products Regulatory Agency (MHRA) is evaluating seven new technologies designed to tackle some of healthcare’s most pressing challenges. The innovations being trialled could slash the time for bowel ******* test results to mere minutes and enable earlier detection of skin ******* and genetic eye diseases. The AI Airlock initiative provides a secure, controlled environment for manufacturers to test these sophisticated systems. This “regulatory sandbox” allows for evaluation of the AI’s effectiveness and limitations, helping to forge a clear path towards eventual regulatory approval and deployment within the health service. The insights gained from this real-world testing will directly inform the MHRA’s future regulations for AI as a medical device and feed into the National Commission into the Regulation of AI in Healthcare. This commission brings together a diverse group of patient advocates, clinicians, regulators, and technology companies to advise the agency. The seven selected technologies include AI-powered clinical note-taking to reduce administrative burdens on doctors, advanced ******* diagnostics, sophisticated eye disease detection, and tools that can summarise a patient’s entire hospital stay or interpret complex blood tests. The ultimate goal is to use AI to support clinicians, helping them to make faster and better-informed care decisions for their patients. Health Innovation Minister Zubir Ahmed said: “The AI airlock programme is a great example of how we can test new innovations thoroughly while still moving at pace, as we seek to deliver on our promise to shift healthcare from analogue to digital. “Through our ten year health plan we will drive for the NHS to be the most AI-enabled healthcare system in the world.” Pioneering a path for safe AI healthcare innovation The evolution of AI presents unique challenges for regulators tasked with ensuring patient care safety. Lawrence Tallon, MHRA Chief Executive, commented: “As the first country to create a dedicated regulatory environment, or ‘sandbox’, specifically for AI medical devices, we’re pioneering solutions to the unique challenges of regulating these emerging healthcare technologies. “The first phase of AI Airlock demonstrated the value of close collaboration between innovators and regulators. I look forward to seeing the results of this new cohort and how their technologies will shape the next generation of safe, effective AI tools in healthcare.” This second phase builds upon the success of the initial pilot. The MHRA has published four reports detailing the key findings from the first group of participants, which included innovative companies like Philips and OncoFlow. Working with the initial cohort, the AI Airlock programme identified several areas for regulatory improvement. These included better methods for validating synthetic data used to train AI models, ensuring the decisions made by an AI are “explainable” to a clinician, and developing novel approaches to tackle emerging risks such as AI “hallucinations,” where a model generates incorrect or nonsensical information. Yinnon Dolev, Gen AI Product Owner at Philips Medical Systems, said: “Participating in the AI Airlock sandbox was a very positive experience. The chance for a R&D representative to impact the regulatory strategy with the regulator is almost unheard of. “The interaction with the team and experts was exceptional. They provided invaluable insights and support, making the entire process smooth and productive. Meeting with the MHRA on a weekly basis throughout the pilot was also a catalyst for meaningful progress expediting our development activities.” A clinician’s perspective of using AI tools for patient care For frontline medical staff, the promise of AI is coupled with a healthy dose of caution. Sir Andrew Goddard, Chairman of the AI Airlock Governance Board and a Consultant Gastroenterologist at Royal Derby Hospital, emphasised the programme’s role in building trust. “Many clinicians, like myself, are keen to see AI find its place in the NHS, but are worried by over-promise on results and lack of reassurance with regards to patient safety,” explained Sir Goddard. “This programme goes a long way to embedding safety and rapid development of these new technologies in our health service.” By bringing innovators and regulators together at an early stage, the AI Airlock will help to deliver the next wave of medical technology to improve patient care, but in a manner that is also safe, reliable, and worthy of the trust of both patients and doctors. See also: NVIDIA GPUs to power Oracle’s next-gen enterprise AI services 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 Expo, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post MHRA fast-tracks next wave of AI tools for patient care appeared first on AI News. View the full article

Oracle and NVIDIA have expanded their partnership to make enterprise AI services more available, powerful, and practical. The announcements, made during Oracle AI World, cover everything from monstrously powerful new hardware to deeply integrated software that aims to put AI at the very core of a company’s data. Ian Buck, VP of Hyperscale and High-Performance Computing at NVIDIA, said: “Through this latest collaboration, Oracle and NVIDIA are marking new frontiers in cutting-edge accelerated computing—streamlining database AI pipelines, speeding data processing, powering enterprise use cases and making inference easier to deploy and scale on OCI.” The headline announcement is the new OCI Zettascale10 computing cluster. This platform is accelerated by NVIDIA GPUs and engineered for the kind of AI training and inference workloads that would make a normal server weep. OCI Zettascale10 promises a mighty 16 zettaflops of peak AI compute performance and is knitted together with NVIDIA’s Spectrum-X Ethernet, a networking fabric designed specifically to stop GPUs from sitting around waiting for data, allowing organisations to scale up to millions of processors efficiently. But raw power is only half the story. The real substance of this partnership lies in the software integrations that aim to weave AI into every layer of a business’s operations. Mahesh Thiagarajan, Executive VP of Oracle Cloud Infrastructure, commented: “OCI Zettascale10 delivers multi‑gigawatt capacity for the most challenging AI workloads with NVIDIA’s next-generation GPU platform. “In addition, the native availability of NVIDIA AI Enterprise on OCI gives our joint customers a leading AI toolset close at hand to OCI’s 200+ cloud services, supporting a long tail of customer innovation.” Giving your Oracle database a brain with AI The foundation of this new strategy is the Oracle AI Database 26ai. For years, the conventional wisdom was to move your data to where the AI models are. Oracle is flipping that on its head, arguing that it’s far more secure and efficient to bring the AI to your data. This latest database release is the embodiment of that “AI for Data” vision. Juan Loaiza, Executive VP of Oracle Database Technologies at Oracle, said: “By architecting AI and data together, Oracle AI Database makes ‘AI for Data’ simple to learn and simple to use. We enable our customers to easily deliver trusted AI insights, innovations, and productivity for all their data, everywhere, including both operational systems and analytic data lakes.” One of the standout features is the ability to run agentic AI workflows inside your database. The AI agents can tackle complex questions by combining your enterprise’s private, sensitive data with public information, all without ever having to move that private data outside your secure environment. This is made possible by features like a Unified Hybrid Vector Search, which lets the AI look for context across all your data types, whether it’s in a relational table, a JSON file, or a spatial map. Oracle is also clearly thinking about the long game with security. The new database implements NIST-approved quantum-resistant algorithms for data both in-flight and at-rest. It’s a defence against “harvest now, decrypt later” attacks, where hackers steal encrypted data today with the hope of breaking it with future quantum computers. Holger Mueller, VP and Principal Analyst at Constellation Research, commented: “Great AI needs great data. With Oracle AI Database 26ai, customers get both. It’s the single place where their business data lives—current, consistent, and secure. And it’s the best place to use AI on that data without moving it. “To help simplify and accelerate AI adoption, AI Database 26ai includes impressive new AI features that go beyond AI Vector Search. A highlight is Oracle’s architecting agentic AI into the database, enabling customers to build, deploy, and manage their own in-database AI agents using a no-code visual platform that includes pre-built agents.” The new database is designed to work with NVIDIA’s toolset. Its programming interfaces can now plug directly into NVIDIA NeMo Retriever, a collection of microservices that handle the complicated plumbing of modern AI for an enterprise. This makes it far easier for developers to implement things like retrieval-augmented generation, or RAG. In simple terms, RAG allows a language model to look up relevant facts in your company documents before it answers a question, making its responses far more accurate and useful. The Oracle Private AI Services Container will also get a GPU-powered boost. This container lets businesses run AI models in their own secure environment. Soon, it will be able to offload the heavy lifting of creating vector embeddings – a core task for AI search – to powerful NVIDIA GPUs using the cuVS library. This promises to slash the time it takes to prepare data for AI applications. Democratising enterprise AI Beyond the database, the partnership aims to simplify the entire AI pipeline. The new Oracle AI Data Platform now includes a built-in NVIDIA GPU option and the NVIDIA RAPIDS Accelerator for Apache Spark. For data scientists and engineers, this is a big deal. It means they can speed up their data processing and machine learning workflows using GPUs, often without having to change a single line of their existing code. All of these tools and capabilities are being consolidated within the Oracle AI Hub. The idea is to give organisations a single place to build, deploy, and manage their AI solutions. From the hub, users can deploy NVIDIA’s NIM microservices – which are like pre-packaged AI skills – through a simple, no-code interface. To lower the barrier to entry even further, the full NVIDIA AI Enterprise software suite is now natively available within the OCI Console. This means that a developer can spin up a GPU instance and enable all the necessary NVIDIA tools with a few clicks, rather than going through a separate procurement process. It’s a small change that makes a big difference in how quickly teams can get started. It’s clear that this collaboration is aimed at solving the real-world challenges businesses face when trying to adopt AI. By bringing the hardware, the data, and the software tools into one cohesive ecosystem, Oracle and NVIDIA are making a case that the era of practical, secure, and scalable enterprise AI has well and truly arrived. See also: Cisco: Only 13% have a solid AI strategy and they’re lapping rivals 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 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 GPUs to power Oracle’s next-gen enterprise AI services appeared first on AI News. View the full article

In a cement plant operated by Conch Group, an agentic AI system built on Huawei infrastructure now predicts the strength of clinker with over 90% accuracy and autonomously adjusts calcination parameters to cut coal consumption by 1%—decisions that previously required human expertise accumulated over decades This exemplifies how Huawei is developing agentic AI systems that move beyond simple command-response interactions toward platforms capable of independent planning, decision-making, and execution. Huawei’s approach to building these agentic AI systems centres on a comprehensive strategy spanning AI infrastructure, foundation models, specialised tools, and agent platforms. Zhang Yuxin, CTO of Huawei Cloud, outlined this framework at the recent Huawei Cloud AI Summit in Shanghai, where over 1,000 leaders from politics, business, and technology examined practical implementations across finance, shipping ports, chemical manufacturing, healthcare, and autonomous driving. The distinction matters because traditional AI applications respond to user commands within fixed processes, while agentic AI systems operate with autonomy that fundamentally changes their role in enterprise operations. Zhang characterised this as “a major shift in applications and compute,” noting that these systems make decisions independently and adapt dynamically, reshaping how computing systems interact and allocate resources. The question for enterprises becomes: how do you build infrastructure and platforms capable of supporting this level of autonomous operation? What do tomatoes and cement have in common? Watch a behind-the-scenes taster of how Huawei & Conch Group use AI to reshape the construction industry! Next up on the intelligent transformation menu: a mouthwatering new era of architecture—smarter, faster, cheaper, greener! pic.twitter.com/hEVIQ0xtUZ — Huawei (@Huawei) August 28, 2025 Infrastructure challenges drive new computing architectures The computational demands of agentic AI systems have exposed limitations in traditional cloud architectures, particularly as foundation model training and inference requirements surge. Huawei Cloud’s response involves CloudMatrix384 supernodes connected through a high-speed MatrixLink network, creating what the company describes as a flexible hybrid compute system combining general-purpose and intelligent compute capabilities. The architecture specifically addresses bottlenecks in Mixture of Experts (MoE) models through expert parallelism inference, which reduces NPU idle time during data transfers. According to the company’s technical specifications, this approach boosts single-PU inference speed 4-5 times compared to other popular models. The system also incorporates memory-centric AI-Native Storage designed for typical AI tasks, aimed at enhancing both training and inference efficiency. ModelBest, a company specialising in general-purpose AI and device intelligence, demonstrated practical applications of this infrastructure. Li Dahai, co-founder and CEO of ModelBest, detailed how their MiniCPM series—spanning foundation models, multi-modal capabilities, and full-modality integration—integrates with Huawei Cloud AI Compute Service to achieve 20% improvements in training energy efficiency and 10% performance gains over industry standards. The MiniCPM models have found applications in automotive systems, smartphones, embodied AI, and AI-enabled personal computers. From foundation models to industry-specific applications The challenge of adapting foundation models for specific industry needs has driven the development of more sophisticated training methodologies. Huawei Cloud’s approach encompasses three key components: a complete data pipeline handling collection through management, a ready-to-use incremental training workflow, and a smart evaluation platform with preset evaluation sets. The incremental training workflow reportedly boosts model performance by 20-30% through automatic adjustment of data and training settings based on core model features and industry-specific objectives. The evaluation platform enables quick setup of systems aligned with industry or company benchmarks, addressing both accuracy and speed requirements. Real-world implementations illustrate the practical application of these methodologies. Shaanxi Cultural Industry Investment Group partnered with Huawei to integrate AI with cultural tourism operations. Huang Yong, Chairman of Shaanxi Cultural Industry Investment Group, explained that using Huawei Cloud’s data-AI convergence platform, the organisation combined diverse cultural tourism data to create comprehensive datasets spanning history, film, and intangible heritage. The partnership established what they term a “trusted national data space for cultural tourism” on Huawei Cloud, enabling applications including asset verification, copyright transaction, enterprise credit enhancement, and creative development. The collaboration produced the Boguan cultural tourism model, which powers AI-driven tools, including a cultural tourism intelligent brain, smart management assistant, intelligent travel assistant, and an AI short video platform. International implementations demonstrate similar patterns. Dubai Municipality worked with Huawei Cloud to integrate foundation models, virtual humans, digital twins, and geographical information systems into urban systems. Mariam Almheiri, CEO of the Building Regulation and Permits Agency at Dubai Municipality, shared how this integration has improved city planning, facility management, and emergency responses. Enterprise-grade agent platforms emerge The distinction between consumer-focused AI agents and enterprise-grade agentic AI systems centres on integration requirements and operational complexity. Enterprise systems must seamlessly integrate into broader workflows, handle complex situations, and meet higher operational standards than consumer applications designed for quick interactions. Huawei Cloud’s Versatile platform addresses this gap by providing infrastructure for businesses to create agents tailored to production needs. The platform combines AI compute, models, data platforms, tools, and ecosystem capabilities to streamline agent development through deployment, release, usage, and management phases. Conch Group’s implementation in cement manufacturing offers specific performance metrics. The company partnered with Huawei to create what they describe as the cement industry’s first AI-powered cement and building materials model. The resulting cement agents predict clinker strength at 3 and 28 days with predictions deviating less than 1 MPa from actual results, representing over 90% accuracy. For cement calcination optimisation, the model suggests key process parameters and operational solutions that cut standard coal usage by 1% compared to class A energy efficiency standards. Xu Yue, Assistant to Conch Cement’s General Manager, noted that the model’s success with quality control, production optimisation, equipment management, and safety establishes groundwork for end-to-end collaboration and decision-making through cement agents, moving the industry “from relying on traditional expertise to being fully driven by data across all processes.” In corporate travel management, Smartcom developed a travel agent using Huawei Cloud Versatile that provides end-to-end smart services across departure, transfers, and flights. Kong Xianghong, CTO of Shenzhen Smartcom and Director of Smartcom Solutions, reported that the system combines travel industry data, company policies, and individual trip histories to generate recommendations. Employees adopt over half of these suggestions and complete bookings in under two minutes. The agent resolves 80% of issues in an average of three interactions through predictive question matching. What’s next for autonomous AI? The implementations discussed at the summit reflect a broader industry trend toward agentic AI systems that operate with increasing autonomy within defined parameters. The technology’s progression from reactive tools to systems capable of planning and executing complex tasks independently represents a fundamental architectural shift in enterprise computing. However, the transition requires substantial infrastructure investments, sophisticated data engineering, and careful integration with existing business processes. The performance metrics from early implementations—whether in manufacturing efficiency gains, urban management improvements, or travel booking optimisation—provide benchmarks for organisations evaluating similar deployments. As agentic AI systems continue to mature, the focus appears to be shifting from technological capability demonstrationsto operational integration challenges, governance frameworks, and measurable business outcomes. The examples from cement manufacturing, cultural tourism, and corporate travel management suggest that practical value emerges when these systems address specific operational pain points rather than serving as general-purpose automation tools. (Photo by AI News ) See also: Huawei details open-source AI development roadmap at Huawei Connect 2025 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 How Huawei is building agentic AI systems that make decisions independently appeared first on AI News. View the full article