Streamers Sue Twitch and Amazon Over AI Training Data Fiasco
A new class-action lawsuit claims that Twitch and Amazon used millions of streamers’ content to train Amazon’s AI models without securing proper consent or offering payment. The complaint was filed in the US District Court for the Northern District of California, with Warren Pandiscia, a Twitch creator from Connecticut, as the lead plaintiff.
Earlier this month, Twitch quietly enabled a “Training for Generative AI” setting on all creator accounts. When toggled on, the setting allows Twitch and its parent company, Amazon, to use a channel’s livestreams, video-on-demand uploads, clips, chat messages, images, and text for AI training purposes. Although Twitch added an opt-out option after public outrage, many users were still upset that the setting was enabled by default.
Central to the lawsuit’s argument is that once content feeds into an AI system, creators cannot “claw back” what the complaint argues is intellectual property. It further claims Twitch and Amazon used creators’ content to train their AI products and sold that data without obtaining the proper licenses, as reported by Courthouse News.
The complaint says that the per-channel opt-out design lets Twitch capture a creator’s content when they appear in or chat on an opted-in channel, even if they opted out on their own channel. This design allegedly makes it impossible to get consent from all parties in captured communications.
Pandiscia’s complaint asks the court to certify a nationwide class of affected users, stop all of the above, and order Twitch and Amazon to identify, segregate, and delete unlawfully obtained data. Neither company has commented on the lawsuit.
