The lawsuit claims that Apple used a dataset consisting of millions of YouTube videos to train an AI model, as revealed in a study published in late 2024. Here are the details.

Amazon and OpenAI Also Accused

As noted by MacRumors, a class action lawsuit proposal filed by Ted Entertainment, Matt Fisher, and Golfholics alleges that Apple downloaded millions of videos by bypassing YouTube's anti-scraping protections.

According to the document, a group of Apple researchers published a study titled STIV: Scalable Text and Image Conditional Video Generation, stating that they trained the video generation model described in the article using a dataset called Panda-70M.

Excerpt from the lawsuit:

The Panda 70M dataset serves as a map or index file identifying specific YouTube videos and clips by URL, video identifier, and timestamp. A single YouTube video can be divided into many clips, each treated as a separate training example. Extracting any clip requires independent access to the source video on YouTube and isolating the specified segment; this creates a separate excessive transition action for each extracted clip.

The plaintiffs claim that their content appeared in the dataset more than 500 times and wish to file a class action on behalf of “everyone in a similar situation.”

Essentially, they argue that the dataset only provides links to videos, but Apple used its access to bypass YouTube's anti-scraping protections to download the underlying content and train AI models.

Therefore, they are requesting a jury trial for all claims and are seeking the following:

  1. Approval of this lawsuit as a class action and appointment of attorneys for the Defendants and Plaintiffs to represent the Class;
  2. A declaration that the Defendant has intentionally bypassed YouTube's copyright protection systems to protect the audiovisual content of the Plaintiffs and Class Members.
  3. Damages (up to the maximum legally permitted per violation), injunctive relief, and attorney's fees and costs under 17 U.S.C. §1203;
  4. Equitable relief necessary to prevent or restrict the infringement of the copyright-protected content of the Plaintiffs and Class Members, under Title 17, Title 28, and/or the Court's inherent authority, as well as requiring the Defendant and its officials, agents, employees, attorneys, directors, successors, assigns, licensees, and all others actively cooperating with them to cease, promote, or contribute to infringing the federal rights of the Plaintiffs or Class Members;
  5. Requesting pre- and post-judgment interest on any monetary award against the Defendant to the fullest extent possible; and
  6. Any other relief that the Court deems just and appropriate.

In addition to Apple, the plaintiffs have also proposed class action lawsuits against Amazon and OpenAI, claiming that both companies used the Panda-70M dataset in their own AI model training processes.

Products Worth Browsing on Amazon

  • David Pogue – 'Apple: The First 50 Years'
  • MacBook Neo
  • Logitech MX Master 4
  • AirPods Pro 3
  • AirTag (2nd Generation) – 4 Pack
  • Apple Watch Series 11
  • Wireless CarPlay Adapter