Both OpenAI and Google have turned to transcribing YouTube videos to further train their AI models, potentially infringing on creators’ copyrights, New York Times Report. The report details how the two tech giants, along with Meta, cut corners to obtain as much data as possible to train their artificial intelligence models.
OpenAI’s Sora just released a psychedelic music video to fan the flames of AI hype
According to reports, OpenAI used the speech recognition tool Whisper to transcribe more than 1 million hours of YouTube videos. It then feeds the record into GPT-4, the powerful artificial intelligence system running the latest version of the ChatGPT chatbot. Google, which owns YouTube, also transcribes YouTube videos to train its AI models.
The two companies’ transcriptions of the videos could infringe on the copyrights of their video creators. Other uses of creator content to train artificial intelligence have prompted copyright and licensing lawsuits.
OpenAI’s use of YouTube videos may also violate Google’s rules, which prohibit the use of its videos in “standalone” applications and “automated means (such as bots, botnets, or scrapers)” that access its videos.
Google spokesman Matt Bryant told the New York Times that the company was not aware of any such use of OpenAI. But the report said Google personnel knew OpenAI was using YouTube videos without authorization but neglected to take action because it was doing the same thing. Google also told the newspaper that it only trains its AI on videos from creators who agree to have their content used in this way.
In July 2023, Google changed its terms of service to allow the use of public online materials such as Google Docs and Google Maps restaurant reviews to further train its artificial intelligence models.