Snap just became the latest tech giant dragged into court over AI training practices. A group of YouTubers with 6.2 million combined subscribers filed a proposed class action lawsuit on Friday, alleging the social media company scraped their videos without permission to power AI features like Imagine Lens. The case adds Snap to a growing list of defendants that already includes Nvidia, Meta, and ByteDance, as content creators push back against what they see as wholesale theft of their work.
Snap is facing a fresh legal battle over how it trained its artificial intelligence models. A group of YouTube content creators filed a proposed class action lawsuit in California federal court on Friday, alleging the company scraped their videos without permission to build AI-powered features that now generate revenue for the platform.
The lawsuit, filed in the U.S. District Court for the Central District of California, centers on Snap's use of the HD-VILA-100M dataset - a massive collection of video-language pairs that was explicitly designed for academic and research purposes only. The plaintiffs claim Snap violated YouTube's terms of service, circumvented technological protections, and ignored licensing restrictions that prohibit commercial use of such datasets.
The case is led by the creators behind the popular h3h3 YouTube channel, which boasts 5.52 million subscribers, along with two smaller golf-focused channels, MrShortGame Golf and Golfoholics. Together, the trio represents roughly 6.2 million collective subscribers whose content allegedly ended up training Snap's AI systems without compensation or consent.
At the heart of the complaint is Snap's Imagine Lens feature, which lets users edit images using text prompts. According to the plaintiffs, this commercial AI tool was built on the backs of their copyrighted video content, extracted through datasets that were never meant to leave the lab. The lawsuit seeks statutory damages and a permanent injunction to halt what the creators characterize as ongoing copyright infringement.
This isn't the first rodeo for these particular YouTubers. They've already filed similar lawsuits against Nvidia, Meta, and ByteDance, according to court documents reported by Law360 and Bloomberg Law. The coordinated legal strategy suggests content creators are getting organized in their fight against what they view as systematic exploitation of their work by deep-pocketed tech companies racing to dominate the AI landscape.
The timing couldn't be more charged. AI companies are under mounting pressure from multiple creative industries over their training data practices. According to the Copyright Alliance, a non-profit advocacy organization, over 70 copyright infringement cases have now been filed against AI companies. The wave of litigation spans publishers, authors, newspapers, artists, and user-generated content platforms, all demanding answers about how their work ended up inside commercial AI systems.
But the legal landscape remains murky. In some cases, judges have sided with tech giants. Last June, a federal judge ruled in favor of Meta in a lawsuit brought by a group of authors over training AI models on copyrighted books. In other instances, AI companies have opted to settle rather than face prolonged litigation - Anthropic paid out $1.5 billion to resolve claims from writers, though the settlement drew criticism for its terms.
For Snap, the lawsuit represents a new front in an already complicated year. The company has been aggressively pushing into AI features to compete with rivals like Meta's Instagram and ByteDance's TikTok. Last October, Snap rolled out its first open-prompt AI Lens for free in the U.S., letting users generate and edit images through text commands. But if the plaintiffs prevail, those features could face significant disruption.
The broader question looming over all these cases is whether existing copyright law can keep pace with AI development. Tech companies argue that training AI models on publicly available data constitutes fair use - a legal doctrine that allows limited use of copyrighted material without permission. Content creators counter that scraping millions of videos or articles to build commercial products goes far beyond what fair use was designed to protect.
Snap declined to comment when reached by TechCrunch. The silence is telling. As more cases wind through the courts, AI companies are walking a tightrope - publicly defending their data collection practices while quietly negotiating licensing deals with major content providers to avoid exactly these kinds of legal headaches.
What happens next could reshape the entire AI industry's approach to training data. If courts consistently rule that using copyrighted material without permission violates intellectual property law, AI companies may be forced to either license content at scale or rely on synthetic and public domain data - both of which present their own technical and financial challenges.
The Snap lawsuit marks an inflection point in the battle over AI training data. With over 70 cases now filed and outcomes ranging from outright dismissals to nine-figure settlements, the legal framework governing how AI companies can use copyrighted content remains unsettled. For content creators like the YouTubers behind h3h3, the fight is about more than money - it's about establishing precedent for an industry that's already racing ahead of the law. As AI features become core revenue drivers for platforms like Snap, expect these courtroom battles to intensify before any clear resolution emerges.
