Twitch Enables Default AI Training on Streamers Content for Amazon
Twitch has implemented a policy allowing its parent company, Amazon, to use streamers’ content to train generative artificial intelligence models by default, requiring creators to manually opt out if they do not wish to participate.
What Happened
Under the new update, recorded video and audio streams from Twitch creators are designated to help build and train Amazon’s generative AI content models. The platform introduced a specific toggle within user channel settings to allow streamers to disable this data usage.
The announcement sparked swift backlash across the Twitch community. During an official live broadcast with nearly 3,000 viewers, Twitch Head of Community Mary Kish and Chief Product Officer Mike Minton addressed community concerns directly. When addressing why the feature was made opt-out instead of opt-in, Minton stated that if the system required creators to opt in voluntarily, nobody would choose to do so.
Questions were also raised regarding whether historical channel content had already been provided to Amazon. Minton acknowledged that he did not know whether Amazon had previously utilized such data for model training.
Key Highlights
- Amazon will use Twitch stream recordings—including extensive video and audio content—to train generative AI models by default.
- Twitch management confirmed that the decision to make the policy opt-out rather than opt-in was driven by the expectation that users would not voluntarily agree to participate.
- Twitch leadership noted uncertainty over whether creator content has already been utilized by Amazon for prior AI training.
- Community representatives stated that providing an opt-out mechanism was a direct response to community pushback against generative AI data harvesting.
- To disable data sharing, creators must go to their channel settings (not the creator dashboard), open the security and privacy tab, locate the “training for generative AI” section, and turn off the toggle.
Why This Matters
Streaming broadcasts provide thousands of hours of vocal and visual recordings, making them highly useful for AI training datasets. However, because creators stream their likenesses and voices for many hours each week, default inclusion has caused concern among users worried about surrendering personal content without realization.
Twitch leadership pointed out that other major technology companies follow similar practices. For instance, Meta utilizes public posts across Facebook and Instagram to train AI models, offering opt-out capabilities in specific regions like the United Kingdom, while requiring users in other locations to set accounts to private to avoid data use.
What to Watch Next
Content creators seeking to prevent Amazon from accessing their streams for AI development must manually review and adjust their channel privacy settings.
Frequently Asked Questions
How can Twitch creators opt out of Amazon AI training?
Creators must navigate to their channel settings rather than the creator dashboard, select the security and privacy tab, scroll down to the “training for generative AI” option, and toggle the setting off.
Why is participation set to opt-out instead of opt-in?
According to Twitch Chief Product Officer Mike Minton, the platform implemented an opt-out structure because users would largely decline to participate if required to opt in actively.
Has Amazon already trained models on past Twitch videos?
Twitch leadership stated during a live community stream that they do not know what Amazon has previously used or not used regarding model training.
Source: TechCrunch, Twitch.