OpenAI and Google reportedly used transcriptions of YouTube videos to train their AI models

[Pelican Press 0]

OpenAI and

This is the hidden content, please

• Sign In
• or
• Sign Up

reportedly used transcriptions of

This is the hidden content, please

• Sign In
• or
• Sign Up

videos to train their AI models

OpenAI and

This is the hidden content, please

• Sign In
• or
• Sign Up

trained their AI models on text transcribed from

This is the hidden content, please

• Sign In
• or
• Sign Up

videos, potentially violating creators’ copyrights, according to

This is the hidden content, please

• Sign In
• or
• Sign Up

. The report, which describes the lengths OpenAI,

This is the hidden content, please

• Sign In
• or
• Sign Up

and Meta have gone to in order to maximize the amount of data they can feed to their AIs, cites numerous people with knowledge of the companies’ practices. It comes just days after

This is the hidden content, please

• Sign In
• or
• Sign Up

CEO Neal Mohan said in an interview with

This is the hidden content, please

• Sign In
• or
• Sign Up

that OpenAI’s alleged use of

This is the hidden content, please

• Sign In
• or
• Sign Up

videos to train its new text-to-video generator, Sora, .

According to the NYT, OpenAI used its Whisper speech recognition tool to transcribe more than one million hours of

This is the hidden content, please

• Sign In
• or
• Sign Up

videos, which were then used to train GPT-4.

This is the hidden content, please

• Sign In
• or
• Sign Up

previously reported that OpenAI had used

This is the hidden content, please

• Sign In
• or
• Sign Up

videos and podcasts to train the two AI systems. OpenAI president Greg Brockman was reportedly among the people on this team. Per

This is the hidden content, please

• Sign In
• or
• Sign Up

’s rules, “unauthorized scraping or downloading of

This is the hidden content, please

• Sign In
• or
• Sign Up

content” is not allowed, Matt Bryant, a spokesperson for

This is the hidden content, please

• Sign In
• or
• Sign Up

, told NYT, also saying that the company was unaware of any such use by OpenAI.

The report, however, claims there were people at

This is the hidden content, please

• Sign In
• or
• Sign Up

who knew but did not take action against OpenAI because

This is the hidden content, please

• Sign In
• or
• Sign Up

was using

This is the hidden content, please

• Sign In
• or
• Sign Up

videos to train its own AI models.

This is the hidden content, please

• Sign In
• or
• Sign Up

told NYT it only does so with videos from creators who have agreed to take part in an experimental program. Engadget has reached out to

This is the hidden content, please

• Sign In
• or
• Sign Up

and OpenAI for comment.

The NYT report also claims

This is the hidden content, please

• Sign In
• or
• Sign Up

tweaked its privacy policy in June 2022 to more broadly cover its use of publicly available content, including

This is the hidden content, please

• Sign In
• or
• Sign Up

Docs and

This is the hidden content, please

• Sign In
• or
• Sign Up

Sheets, to train its AI models and products. Bryant told NYT that this is only done with the permission of users who opt into

This is the hidden content, please

• Sign In
• or
• Sign Up

’s experimental features, and that the company “did not start training on additional types of data based on this language change.”

This is the hidden content, please

• Sign In
• or
• Sign Up

news, gear,

This is the hidden content, please

• Sign In
• or
• Sign Up

, openai,

This is the hidden content, please

• Sign In
• or
• Sign Up

#OpenAI #

This is the hidden content, please

• Sign In
• or
• Sign Up

#reportedly #transcriptions #

This is the hidden content, please

• Sign In
• or
• Sign Up

#videos #train #models

This is the hidden content, please

• Sign In
• or
• Sign Up