Apple is pushing back against a class-action lawsuit from several YouTube creators who claim the company unlawfully scraped their
videos to help train AI models. The case is another entry in the fast-growing legal pileup over generative AI training data, where tech giants keep arguing that public web content is fair game, while creators increasingly ask who gave them permission to take it in the first place.
The lawsuit was filed in April in the U.S. District Court for the Northern District of California by Ted Entertainment, Matt Fisher, and Golfholics, the owners behind h3h3Productions, MrShortGame Golf, and Golfholics. The creators allege that Apple bypassed YouTube’s anti-scraping protections to access and download copyrighted videos at scale. Their complaint frames the alleged scraping as part of a larger AI land grab, one where creator-made content gets fed into powerful commercial systems without payment, licensing, or much in the way of a friendly knock at the door.
Apple’s newly filed motion to dismiss takes a very different view. In Apple’s telling, the lawsuit tries to turn YouTube’s rules against scraping into a DMCA access-control claim, and that is where the company says the whole thing falls apart. Apple argues that the creators uploaded their videos to YouTube, where the public could watch them freely. “No password. No payment. No lock. No key.” That phrasing appears directly in Apple’s filing, and it neatly sums up the company’s defense.
The key legal fight centers on Section 1201(a) of the Digital Millennium Copyright Act, which targets circumvention of technological measures that control access to copyrighted works. Apple says YouTube’s alleged technical barriers may limit how someone uses or downloads videos, but they do not block access to the videos themselves. In plain English, Apple is arguing that watching a video through a public YouTube page is not the same thing as cracking open a locked digital vault.
Apple also leans on the distinction between access and use. The company compares true access controls to a black box that prevents anyone from seeing what is inside. Put another way, Apple is treating YouTube less like a locked vault and more like something visible to the public, even if YouTube has rules or tools meant to limit downloading.
MacRumors first reported on Apple’s response, noting that the same YouTube channel owners have filed similar lawsuits against Meta, NVIDIA, ByteDance, and Snap. That broader litigation push makes clear that this is not just an Apple problem. It is an AI industry problem, and it is not going away because one company says “public means public” with a lawyerly flourish
That fight over what counts as
fair game for AI training is not new. In 2024, Microsoft AI CEO Mustafa Suleyman told
CNBC’s Andrew Ross Sorkin at the Aspen Ideas Festival that content on the open web had long been treated as fair use, or “freeware,” unless publishers explicitly told crawlers to stay away. The comment landed badly with creators and publishers, because it sounded less like a legal theory and more like the AI industry saying the internet had already agreed to be its buffet.
For now, Apple wants the court to dismiss the complaint with prejudice, or alternatively strike parts of the requested relief tied to alleged copyright infringement rather than circumvention. No judge has ruled on the motion yet, and the case has not been certified as a class action.
Still, this one is worth watching. A win for Apple could strengthen the argument that publicly available YouTube videos are not protected by the DMCA’s anti-circumvention rules in the way these creators claim. A win for the YouTubers could make AI companies rethink how casually they scoop up creator content from the open web. Either way, the
AI training-data fight just got another big-name defendant swinging back.