Class Action Lawsuit: Content Creators Claim Snap Wrongfully ‘Scrapes’ Content From YouTube to Train AI

New to ClassAction.org? Read our Newswire Disclaimer

A proposed class action lawsuit alleges that Snap, owner and operator of Snapchat, defies YouTube’s measures designed to protect the intellectual property of its creators by illegally downloading millions of videos off the platform to train its generative AI model.

Get class action lawsuit and class action settlement news sent to your inbox – sign up for ClassAction.org’s free weekly newsletter.

According to the 22-page lawsuit, Snap has fraudulently downloaded the visual and audio files of millions of YouTube videos via backend automated tools in a process known as scraping. In doing so, the complaint argues, the technology giant intends to improve the features on its social media app, Snapchat, so its users can “create, edit and recreate” their pictures by prompting the generative artificial intelligence model.

The scraped YouTube videos, according to the case, are deemed content creators’ intellectual property and are safeguarded from unauthorized access under YouTube’s Terms of Service, along with the platform’s technological protection measures, regardless of copyright registration.

“YouTube’s Terms of Service expressly prohibit scraping, unauthorized downloading, bulk extraction, or other forms of data mining of audiovisual content except through expressly permitted features or licensed [application programming interfaces],” the complaint states. “These contractual restrictions operate together with [technological protection measures] to prevent unlicensed access to creators’ videos.”

However, despite these assurances, Snap has allegedly utilized a video-downloading program and virtual machines that go beyond the “ordinary” use of a site visitor to access datasets—called HD-VILA-100M and Panda-70M—that contain detailed information about millions of YouTube videos. Per the case, these tools operate by rotating IP addresses to avoid detection and circumvent protection measures, which the suit contends is in violation of the Digital Millennium Copyright Act.

For Snap’s AI model, as is the case with all other AI models, the more data that can be accessed and digested in training stages, “the better the AI product” will be, the complaint adds, noting that Snap needed to obtain “vast amounts of data” from various sources to develop its generative AI.

Related Reading: Class Action Lawsuit Alleges Apple Illegally Uses Copyrighted Works for AI Training

“Defendant has at all times intended to create a well-trained generative AI product … that will give Defendant an advantage in competition against its competitors in technology and social media spheres, and that Defendant can leverage for commercial purposes in order to increase its own profitability,” the lawsuit alleges.

The plaintiffs are a group of popular content creators behind the channels “h3h3 Productions,” “H3 Podcast Highlights,” “Mr. ShortGame Golf” and “Golfholics” who claim that a number of their videos were scraped by Snap for the purpose of AI training without authorization. According to the complaint, the plaintiffs collectively have hundreds of distinct videos in the HD-VILA-100M and Panda-70M datasets.

“[Snap]’s actions were not only unlawful, but an unconscionable attack on the community of content creators whose content is used to fuel the multi-trillion-dollar generative AI industry without any compensation,” the case summarizes.

The Snap AI class action lawsuit seeks to represent all individuals in the United States who uploaded original videos to YouTube that were partially or entirely included in the HD-VILA-100M and Panda-70M datasets scraped and downloaded by Snap.

Check out ClassAction.org’s lawsuit list for the latest open class actions.