A proposed class action lawsuit was filed Thursday in the U.S. District Court for the Northern District of California against Twitch Interactive Inc. and Amazon.com Inc. by Twitch streamer Warren Pandiscia and others. The complaint alleges the companies unlawfully used streamers’ videos without permission to train Amazon’s generative AI systems, violating implied contracts and state unfair competition laws, according to court documents.
The lawsuit, filed Thursday in the U.S. District Court for the Northern District of California, alleges that Twitch and Amazon “unlawfully copied millions of videos” from creators’ streams to develop Amazon’s generative AI models without negotiating licenses or obtaining consent, according to court documents. Lead plaintiff Warren Pandiscia, a Connecticut-based Twitch creator with more than 900 followers, represents a proposed class of Twitch streamers whose content was allegedly used in the AI training process.
The complaint claims that Twitch and Amazon harvested streamers’ video content—including live streams, archived videos on demand (VODs), clips, chat logs, and channel images—starting as early as 2024 during an internal prototyping phase.
Plaintiffs say this data collection occurred without permission or compensation and continued until Twitch introduced an opt-out setting in mid-August 2026. Twitch publicly announced the “Training for Generative AI” setting around August 12–13, 2026, which is enabled by default, allowing Amazon to use eligible channel content for AI training unless streamers actively disable the feature, according to Twitch’s documentation and reporting by Courthouse News and AI Rights Brief.
The complaint argues that because AI models cannot be “un-trained,” creators cannot recover intellectual property already copied and used, even if they opt out later. Plaintiffs assert that Twitch and Amazon’s conduct violates implied contracts with streamers and state unfair competition laws, depriving creators of control over their works, economic value, and property rights. The lawsuit seeks injunctive relief to halt the use of streamers’ content for AI training without authorization, as well as damages, restitution, and disgorgement of profits allegedly earned from the AI models.
According to the complaint, Amazon sought AI training data “at an unprecedented rate” and therefore relied on Twitch streams instead of obtaining lawful licenses or creator consent. The lawsuit describes the defendants’ AI data pipeline as enrolling “every creator” by default through a setting activated on August 12, 2026. Twitch has stated it does not directly train AI models but permits Amazon to use Twitch content for generative AI training under the default-on setting. The opt-out process requires streamers to navigate to Settings → Security and Privacy and toggle off the “Training for Generative AI” option, which is located near the bottom of the menu.
The filing also highlights that Twitch confirmed in August 2026 that it uses creators’ content for AI training and updated its Terms of Service and Privacy Policy to reflect this practice. Forbes reported that Twitch internally acknowledged “nobody would opt in” to AI training, which influenced the decision to make the setting default-on and require opt-outs rather than opt-ins. Critics have noted that the setting is buried in privacy menus, making it unlikely many streamers would find and disable it without explicit guidance.
The lawsuit follows significant backlash among Twitch streamers and viewers after the opt-out policy was announced. Media outlets including BBC and NPR described the reaction as “furious streamer backlash,” with creators criticizing the default-on approach and lack of prior notice. The complaint references Twitch’s August confirmation and the ensuing controversy as part of the context for the legal action.
The case, captioned Pandiscia v. Twitch Interactive, Inc. and Amazon.com, Inc., remains at an early stage as a proposed class action. Class certification has not yet been granted. The Northern District of California is a common venue for technology-related class actions, and the complaint focuses specifically on Amazon’s use of Twitch streamers’ content to train generative AI, rather than broader platform issues.
Plaintiffs allege that Twitch and Amazon engaged in deceptive and unfair business practices in violation of state consumer protection laws. The complaint states that the defendants’ systems “are incapable of obtaining the consent of all parties to the communications they capture,” referring to streams and associated content. The lawsuit seeks to hold Twitch and Amazon accountable for what it describes as unauthorized use of creators’ intellectual property to develop AI technology.
The outcome of the case could have implications for how online content platforms handle user-generated material in the context of AI training. The lawsuit adds to ongoing legal and public debates over data use, consent, and compensation in the rapidly evolving field of generative artificial intelligence.
Comments are closed.