Amazon and Twitch Face Lawsuit Over Using Livestreams To Train AI

Amazon and its live-streaming subsidiary Twitch are facing a proposed class action lawsuit over allegations that they harvested video content from millions of creators to train generative artificial intelligence models without securing permission or paying licensing fees.

The 37-page complaint was filed in the U.S. District Court for the Northern District of California by Connecticut-based streamer Warren Pandiscia, who streams to roughly 1,000 followers. Representing creators across the platform, the suit alleges that Amazon breached user contracts and violated state unfair competition laws by quietly using streams, chat logs, clips, and photos for AI development for at least two years before disclosing the practice.

The legal action follows intense pushback from the Twitch community after the platform announced in August that creator content would be used to train Amazon's generative AI models with the training toggle set to "on" by default. Streamers expressed outrage that they were forced to manually hunt through dashboard settings to opt out.

Defending the decision to make the feature opt-out rather than opt-in, Twitch Chief Product Officer Mike Minton previously offered a candid explanation to the community:

"If it was opt in, nobody would opt in. So it's going to be on by default, and almost every content service in the world is on by default."

"The thing we're doing here that is unique is different is respecting your wishes to opt out of model training. I know this is not a fan favourite, I know it's very upsetting to the community, but this is where we are."

Minton also acknowledged uncertainty regarding historical data collection, noting he was unsure what Amazon "has done in terms of model training and what they've used and not used", while acknowledging that stream footage had previously been used "in a prototyping capacity."

The lawsuit highlights several loopholes in Twitch's new privacy controls. Because the opt-out mechanism applies to individual channel settings rather than overall user accounts, a creator who disables the setting can still have their likeness and voice scraped if they appear as a guest on another channel where the feature remains active. Furthermore, Twitch confirmed in an official social media post that opting out does not apply retroactively to content harvested prior to August 12th.

Addressing Amazon's commercial motives, Pandiscia’s complaint states:

“Because Amazon AI products are commercialized, Amazon had an overwhelming incentive to acquire training data on an unprecedented scale. Rather than negotiate for lawful licenses or seek permission, defendants accessed the Twitch streams and videos to utilize them as a massive dataset necessary to fuel Amazon’s AI products,”

With streamers producing over 215 million hours of content in early 2026 alone, the lawsuit seeks monetary damages for affected creators and a court order blocking Amazon from continuing to scrape stream data without explicit consent. Neither Amazon nor Twitch has publicly commented on the pending litigation.