Twitch will use streamers' content for Amazon's AI training by default
Twitch has introduced a new setting letting creators opt out of having their content used to train Amazon's generative AI models, but users are enrolled by default. The move has sparked strong backlash from the streaming community.
Twitch, owned by Amazon, has begun using creators' stream recordings to help train its parent company's generative AI models. The new setting is switched on by default for all streamers, meaning users must actively take action to exclude themselves from this data use.
The announcement triggered significant anger among the community, since Amazon considers the thousands of hours of audio and video that streamers generate weekly to be especially valuable training material. Many users expressed concern that their voices and likenesses could be used without clear consent, given that the change follows an opt-out rather than opt-in model.
Company's explanation
During an official Twitch livestream watched by nearly 3,000 viewers, Head of Community Mary Kish and Chief Product Officer Mike Minton acknowledged the decision was deliberate. Minton admitted plainly that if the setting required users to actively opt in, almost no one would do so.
Some streamers asked during the broadcast whether their past recordings had already been used for training, but Minton said he did not know the answer, as he was not informed about what Amazon had already done with that data.
Kish pointed out that Twitch is not alone in this practice — Meta also uses public content from Facebook and Instagram to train its AI models, and for most users the only way to opt out is to make their account private, which is often impractical for creators who monetize their presence.
To opt out of AI training, Twitch users need to go to their channel settings, select the security and privacy tab, and turn off the "training for generative AI" option.
