Thousand streamers launch class action against Twitch over AI training use of content

Twitch is facing a class-action lawsuit from over a thousand streamers who claim the platform is using their live streams, clips, chat logs, and photos to train Amazon's generative AI models without consent or compensation. Warren Pandiscia, a streamer with approximately a thousand followers, filed a 37-page complaint in the Northern District Court of California, alleging that Twitch violated the creators' implied agreements and state unfair competition laws. The lawsuit centers on Twitch's opt-out mechanism, which is described as effectively useless. The platform recently announced that it would use generative AI within its service, with the feature enabled by default and requiring users to manually disable it in settings. Minton, Twitch's Chief Product Officer, stated that if the feature were truly optional, no one would use it, hence the default setting. Streamers are particularly concerned that the opt-out option is not retroactive and is tied to individual channels rather than accounts, meaning their content can still be used for AI training if they appear in a live stream from another channel that hasn't opted out. The complaint alleges that Twitch and Amazon had already been using streamers' content as early as 2024, long before any opt-out mechanism was introduced. Pandiscia stated that he opted out as soon as the setting became available and had no prior reason to believe his content was being used in this manner. The lawsuit highlights the growing conflict between the creator economy and AI training, as user agreements that include 'default consent' are seen as eroding individuals' control over their own content.