Earlier this week, an investigation detailed that Apple and other tech giants had used YouTube captions to train their AI models. This included more than 170,000 videos from the likes of MKBHD and Mr. Beast. Apple then used this dataset to train its open-source OpenELM models, which it released in April.
Apple has now confirmed that 9to5Macbut that OpenELM does not power any of its AI or machine learning features, including Apple Intelligence.
Apple says it created the OpenELM model as a way to contribute to the research community and encourage the development of open-source large language models. In the past, Apple researchers have described OpenELM as a “state-of-the-art open language model.”
According to Apple, OpenELM was created for research purposes only, not to be used to power any of Apple’s Intelligence features. The model has been published open source and is widely available, including on Apple’s Machine Learning Research website.
Since OpenELM is not used as part of Apple Intelligence, this means that the “YouTube Subtitles” dataset is not used to power Apple Intelligence. In the past, Apple has said that Apple Intelligence models were trained “on licensed data, including data selected to improve specific features, as well as publicly available data collected by our web crawler.”
Finally, Apple lets me know that it has no plans to build new versions of the OpenELM model.
If Wired Earlier this week, companies like Apple, Anthropic, and NVIDIA all reported that they were using this “YouTube Subtitles” dataset to train their AI models. This dataset is part of a larger collection called “The Pile,” from the nonprofit EleutherAI.
FTC: We use income generating automatic affiliate links. More.
