Apple faces accusations of using YouTube videos for AI training without permission

A recent investigation shows that Apple used YouTube videos without the creators’ permission to train its AI models, including videos from famous YouTubers.

Apple faces criticism for unauthorized use of YouTube videos to train its AI models. The reports show that Apple has joined other tech giants such as OpenAI, Meta and Google that have previously been accused of using videos to train AI models without the creators’ consent.

According to a report from Wired, the videos were downloaded by third parties and then used to train large language models (LLMs). The practice included more than 170,000 videos, including content from well-known YouTubers such as MKBHD, Jimmy Kimmel, PewDiePie and MrBeast. The investigation also revealed that subtitles from these videos were also used, which is against YouTube’s rules for independent apps and automated access to their content.

Apple Intelligence

Furthermore, Proof News found that some of the most prominent AI companies, including Apple, Anthropic, Nvidia and Salesforce, used material from thousands of YouTube videos to train their models. Apple reportedly used a third-party dataset, known as Pile, to train its OpenELM model, which launched in April. The data compilation, which Apple said was originally intended for academic and educational purposes, was misused to train commercial AIs. Apple has yet to respond to these concerns, which only adds to the uncertainty and criticism surrounding the issue.

Source: pctuning.cz