Wired recently revealed that Apple (and several other companies) were using YouTube video transcripts without the creators’ consent to train their AI models. Apple has now responded to those allegations.
Apple representatives did not deny in a press release that such a situation had taken place. However, they noted that the controversial dataset, which contains transcripts of over 170,000 videos taken without the consent of the creators from almost 50,000 different YouTube channels, was used exclusively during OpenELM training. This language model, presented by Apple in April this year, was created for research purposes only and will not be used in any products. The company representatives also emphasize that this project is no longer being developed.
Apple claims that the AI used in its new operating systems was trained only on licensed, fully legal content, without violating any copyrights.
Source: 9to5Mac
Source: myapple.pl
