A proposed class action lawsuit has been filed against Twitch and its parent company Amazon, accusing the platforms of using content from creators to train generative artificial intelligence without permission or compensation.
Filed in the Northern District of California by streamer Warren Pandiscia, the 37-page legal complaint alleges that the companies unethically copied broadcast videos, chat logs, clips, and images. The suit contends that this practice violates implied user agreements alongside California unfair competition legislation, asserting that unauthorised data harvesting occurred as early as 2024 prior to any public disclosures.
Twitch publicly acknowledged earlier this month that creator material was being used to build artificial intelligence models for Amazon. The announcement triggered immediate backlash from users after it was revealed that AI training is enabled by default. Channel owners must manually adjust settings to opt out, though the exclusion does not apply retroactively or prevent a streamer's appearance on another channel from being scraped.
Responding to creator dissatisfaction, Twitch Chief Product Officer Mike Minton stated that a default opt-out mechanism was necessary, claiming that an opt-in requirement would yield low participation. Neither Twitch nor Amazon, which acquired the livestreaming platform in 2014 for roughly $970 million, has issued a public response to the suit.
About the Author
