Three YouTube channel operators, including the creators behind h3h3Productions, have filed a lawsuit against Apple over claims that the company used their videos without permission to train its artificial intelligence models. The case centers on how video content hosted on platforms like YouTube may have been collected and incorporated into datasets used for building video-focused AI systems.
The complaint argues that Apple accessed large volumes of YouTube videos by bypassing protections designed to prevent automated scraping. According to the filing, the plaintiffs allege that Apple used automated tools to download content at scale while avoiding YouTube’s safeguards, then used that material in datasets for training internal AI models. The channels involved include h3h3Productions, along with its associated H3 Podcast channels, as well as MrShortGame Golf and Golfholics.
Filed in a federal court in California, the lawsuit alleges violations of the Digital Millennium Copyright Act. The core argument is that Apple circumvented technological measures that protect copyrighted works on YouTube, and then used the collected content in a commercial context tied to AI development. The plaintiffs claim that this process involved millions of videos across the platform, including content uploaded by their own channels.
A portion of the complaint points to Apple’s own research disclosures. The plaintiffs reference published material from Apple researchers that discusses training video AI models using large-scale datasets, including Panda-70M. They argue that datasets of this scale may include content sourced from public platforms such as YouTube, and that their videos were among those included in the training data.
The channels involved represent established creator networks. h3h3Productions, created by Ethan and Hila Klein, operates across multiple channels with a combined audience in the millions. MrShortGame Golf and Golfholics focus on golf-related content and also maintain significant subscriber bases. The lawsuit claims that specific videos from these channels were included in the broader pool of content allegedly used for AI training.
At the center of the legal argument is the method of data collection. The plaintiffs claim that Apple relied on systems designed to avoid YouTube’s protections against scraping, including the use of automated downloading tools and techniques that rotate network identities to evade detection. This aspect of the case focuses on how the data was obtained, rather than only how it was used afterward.
The filing seeks monetary damages as well as a court order preventing Apple from continuing the alleged practice of sourcing YouTube content for AI training. It also requests class action certification, which would allow other creators whose content may have been included in similar datasets to potentially be represented in the case.
The lawsuit adds further scrutiny to how publicly available content is treated when used in machine learning workflows, especially when platform protections are involved. Questions around consent, attribution, and compensation continue to surface as more datasets used in AI development are examined in legal settings.
(via MacRumors)
