Apple says it doesn’t use stolen YouTube video content to train its AI or power its devices‘ machine learning features.
a report by Wired revealed that Apple, NVIDIA, and other tech companies were using hundreds of thousands of YouTube videos to train their AIs. The YouTube Subtitles dataset had video transcripts from education and online learning channels, from MIT and Harvard to Mr. Beast, MKBHD, and PewDiePie.
After the controversy, Apple confirmed to 9to5Mac that its open-source OpenELM models used this dataset. However, Cupertino claims they don’t power AI, machine learning, or Apple Intelligence features.
According to the publication, Apple “created the OpenELM model as a way of contributing to the research community and advancing open source large language model development (…).
