Apple Argues Public Videos Need No Permission For AI Training

Apple’s legal team is pushing to end a copyright lawsuit early, and the argument it is making could matter well beyond this particular case. Three YouTube channels, h3h3Productions, MrShortGame Golf, and Golfholics, sued Apple earlier this year claiming the company scraped their videos without permission to train its AI models, bypassing YouTube’s protections against unauthorized downloading in the process.

Apple’s dismissal request doesn’t dispute the scraping so much as challenge whether it was illegal under the DMCA. The company’s position is that because the videos were publicly viewable without a password, payment, or any restricted access, YouTube’s technical measures did not constitute the kind of protected access controls the DMCA is designed to cover.

That framing is doing a lot of work. If a court accepts it, the argument would effectively mean that content posted publicly online carries weaker copyright protections against AI training pipelines, regardless of a platform’s terms of service or download restrictions. The same channels also filed parallel lawsuits against Meta, Nvidia, ByteDance, and Snap, so the legal question here has industry-wide implications.

A Crowded Queue of AI Ambitions

Apple has been building toward more capable on-device and cloud AI features, with Siri still in development and significant AI hardware investments expected in upcoming devices. The training data underlying those features is now the subject of active litigation across multiple fronts.

The court has not yet ruled on Apple’s dismissal request. If the case survives this stage, it moves into discovery, which is where the details of how Apple actually accessed and used the video content would likely surface. That is probably the outcome Apple is most motivated to avoid.