> In an email to Mashable, Apple said that its open-source language model, OpenELM, indeed used the dataset, but not in the way some may be thinking. The OpenELM project is a part of Apple's ongoing effort to benefit the broader research community. In other words, according to Apple, the OpenELM model was created for research purposes only and will not underpin any of Apple's machine learning-powered hardware or AI services, including Apple Intelligence.

I guess they're just doing research for fun then.

The claims don't even make sense to me.

Apple uses task-centric on-device models and then reach out to OpenAI for the full LLM experience.

YouTube dataset doesn't seem useful for this goal.

Wouldn't it be more accurate to say Pile did it.

I've heard this response and it still doesn't make sense to me. If I get caught downloading a movie, it's not a defence to point to the ripper and say "I didn't violate the copyright, they did!"

I don't see how it's a valid defence at all.

TL;DR: They used it for OpenELM models, not Apple Intelligence models.

There's no clarification of what exactly OpenElm will be used for, just "for the betterment of humanity" or similar corpspeak.

I think the article said it wouldn't be used for anything. There are no intentions to use it for anything further besides releasing it to the research community, which has been completed.