A class action lawsuit has been filed against Twitch and its parent company Amazon in the U.S. District Court for the Northern District of California on August 20, 2026. The suit, led by Connecticut-based creator Warren Pandiscia who has over 900 followers on Twitch, accuses the companies of using creators' streams to train generative artificial intelligence models without permission or compensation. The complaint alleges claims of breach of implied and express contract, unjust enrichment, and unfair business practices. The lawsuit characterizes the resulting harm to creators as permanent, stating that "Content creators such as plaintiff and the class members will never be able to claw back the intellectual property unlawfully copied and used by defendants to train Amazon’s generative AI." According to court documents, the companies initiated the scraping of creator content for AI training purposes as early as 2024. This timeline predates the introduction of an opt-out setting by nearly two years, leaving creators unaware that their broadcasts were being used for Amazon's machine learning systems. The lawsuit points directly to past public statements by Twitch executives. Mike Minton, who previously served as Twitch's Chief Monetization Officer and currently holds the position of Chief Product Officer, acknowledged in 2024 that the platform's content was already being used for AI development in what he described as a "prototyping" capacity. The dispute escalated after Twitch confirmed its practice of using creator content to train Amazon's AI models. In response to backlash, the company introduced an opt-out toggle on August 12, but configured the system so that all channels were opted in by default. Addressing the decision during an August 12 stream, Minton defended the approach, stating: "If it was opt-in, nobody would opt-in. That's honestly the answer." The lawsuit also identifies a loophole in Twitch's opt-out system. Because the setting is applied on a per-channel basis rather than tied to individual user identities, a creator who has disabled AI training on their own channel can still have their voice, likeness, and text captured if they appear on or participate in a chat room hosted by a channel that remains opted in. Pandiscia stated that he disabled the setting immediately upon learning of its existence and would have acted sooner had the companies disclosed the practice. Through the class action, the plaintiffs seek injunctive relief to halt the practice, along with damages, restitution, and the disgorgement of profits Amazon and Twitch generated from using the contested training data. Content creators broadcasting from Somali cities face these default data extraction policies with limited options, as Amazon's localized support and legal representation do not extend to the region. A streamer operating out of Hargeisa or Mogadishu who frequently collaborates on international channels remains exposed to the cross-channel scraping loophole regardless of their local account settings.