Amazon Faces Class Action Over Twitch AI Training Practices

Amazon sued over Twitch AI training

Amazon is facing a class action lawsuit alleging it improperly used content from Twitch streamers to train artificial intelligence models without obtaining adequate permission or providing compensation.

The legal action, brought by Connecticut-based streamer Warren Pandiscia on behalf of millions of Twitch users, claims Amazon breached its contract with creators by repurposing their broadcast videos as AI training data. Streamers produced more than 215 million hours of content in just the first months of 2025, according to the complaint.

Neither Twitch nor Amazon has commented on the lawsuit.

Why it matters

This case highlights growing tensions over how tech companies acquire training data for AI systems. As platforms race to develop competitive AI capabilities, the question of whether user-generated content can be exploited for commercial AI development without explicit consent or payment is becoming a flashpoint. The outcome could set precedents for how streaming platforms and social media companies handle creator content in the AI era.

Opt-out system faces criticism

Amazon announced in August that it would use Twitch content for AI training, offering users an opt-out mechanism. The move immediately drew backlash from the creator community.

To disable the feature, streamers must navigate to their Streamer Dashboard settings, select the Security and Privacy tab, and toggle off "training for Generative AI" near the bottom of the options menu. However, the opt-out only applies to individual streams, meaning creators can still have their content used for training when they appear in other streamers' broadcasts where the feature remains enabled.

Twitch's FAQ documentation indicates that audio from streams may be used to "refine models that create speech to text," with the stated goal of improving automatic subtitles on both Twitch streams and Amazon videos.

Uncertainty over historical data use

The timeline of Amazon's data collection practices remains unclear. Twitch chief product officer Mike Minton previously acknowledged he did not know whether user data had been scraped for training purposes before the opt-out option became available. He stated uncertainty about "what Amazon has done in terms of model training and what they've used and not used."

The lawsuit seeks both financial damages for affected streamers and a court order preventing Amazon from continuing the alleged practice.

Amazon acquired Twitch for nearly $1 billion in 2014 and has since made AI development a central priority as it competes with Google, Meta, and other technology giants for dominance in the artificial intelligence space.

These details were first reported by BBC News.