Twitch faces backlash over Amazon training AI on its users’ livestreams

Twitch Faces Creator Backlash Over Default AI Training Policy

Default Opt-In Policy Sparks Streamer Outcry

Streaming platform Twitch has ignited a firestorm within its community by confirming that user content is now being harvested to train generative AI content models across Amazon. For Amazon, these stream recordings represent a significant resource, offering thousands of hours of audio and video content. Rumors of an AI training program had circulated for at least a few weeks—notably including a post by streamer Zach Bussey on Bluesky back in July—but the policy was officially acknowledged when the company introduced a setting on X, formerly Twitter, that allows users to disable the feature.

Because the feature applies automatically, users found themselves opted in without explicit consent. Reaction on social media and community forums was immediate and hostile. A Twitch User Voice thread on the topic, which originally appeared in May, exploded to 228 pages and nearly 14,000 votes in favor of making all AI features optional and opt-out by default. Responses to the company’s official announcement on X were predominantly composed of instructions on how to disable the training or enthusiastic use of profanity.

Executives Defend the Opt-In Default During Live Stream

During a broadcast on the official Twitch channel, Head of Community Mary Kish and Chief Product Officer Mike Minton addressed an audience of nearly 3,000 aggrieved users, many of whom were posting anti-AI sentiments in the live chat. When questioned about why the program was not configured as opt-in, Minton offered a candid explanation regarding the company’s strategy.

If this was opt-in, nobody would opt in. That’s honestly the answer, Minton stated. He further explained that AI training collection has been turned on by default, similar to the standard practice on most other content services, but claimed the platform is now “respecting” the wishes of those who prefer to opt out.

Kish addressed the reception of the news directly, noting, We don't expect you to be happy or excited about this. I don't expect anyone to react to this favorably.

Uncertainty Over Past Data Harvesting

Streamers have expressed significant unease regarding the scope of the data usage, fearing their hours of recorded live content and voice data could be absorbed into Amazon’s models without their awareness. This concern is heightened by the uncertainty surrounding when the data collection actually began. When asked directly if Amazon had already utilized streamer content for training, Minton admitted that he was unsure of the extent of the activity. I don't know what Amazon in this case has done in terms of model training, and what they've used and not used, Minton said. I'm not responsible for Amazon's training efforts.

The controversy highlights the tension between Amazon’s heavy investment in AI development—as it seeks to compete with technology giants like Google and Meta—and the creators who populate its streaming platform. High-profile streamers have taken to social media to voice their opposition, with user Legundo writing on X: EVERYONE make sure to turn this off and tell Twitch we don't want this crap!

Twitch remains a major livestreaming platform, particularly popular among youths and gamers, with top content creators boasting tens of millions of followers. As the platform navigates this reaction, the core of the dispute remains the balance between Amazon’s desire to leverage massive amounts of data for its AI services and the creators’ demand for control over their recorded personal content.