Snap sued by YouTubers over alleged use of videos to train AI models

A group of prominent YouTubers has expanded its legal battle against artificial intelligence developers by filing a lawsuit against Snap, alleging the company used their videos without consent to train AI systems powering features on Snapchat.

The proposed class action was filed on Friday in the U.S. District Court for the Central District of California by internet content creators who collectively reach more than 6.2 million subscribers across three YouTube channels. The plaintiffs claim Snap trained its AI models using their video content, which was allegedly scraped from YouTube in violation of platform rules and copyright protections.

Also read: Elon Musk slams WhatsApp security as Meta faces lawsuit over encryption claims

According to the complaint, Snap used large-scale video-language datasets, including one known as HD-VILA-100M, that were created strictly for academic and research purposes. The creators allege that Snap repurposed these datasets for commercial AI development, circumventing YouTube’s technological safeguards, terms of service, and licensing restrictions that prohibit such use.

The lawsuit claims that the scraped content was used to support Snapchat AI features such as the “Imagine Lens,” a tool that allows users to edit images using text-based prompts. The plaintiffs argue that their copyrighted videos were exploited to generate commercial value without authorization or compensation.

Also read: No more AI friends as Meta blocks teen access to AI characters

The case is being led by the creators behind the h3h3 Productions YouTube channel, which has more than 5.5 million subscribers, along with creators from the MrShortGame Golf and Golfholics channels. The plaintiffs are seeking statutory damages and a permanent injunction to block any future use of their content in AI training.

Snap is the latest company to be added to a growing list of defendants in similar lawsuits brought by the same group of creators. The plaintiffs have previously sued Nvidia, Meta, and ByteDance, alleging comparable misuse of copyrighted video content for AI development.

The case adds to a broader legal reckoning over how AI companies source training data. Content creators, publishers, artists, authors, and news organizations have increasingly challenged AI firms over the use of copyrighted material. According to the nonprofit Copyright Alliance, more than 70 copyright infringement lawsuits have been filed against AI companies to date.

Courts have delivered mixed outcomes so far. In some instances, judges have ruled in favor of technology companies, while in others, AI developers have reached settlements with rights holders. Many of the cases remain unresolved, leaving key legal questions about copyright, fair use, and AI training practices still open.