Twitch adds opt-out toggle for Amazon AI training

Twitch has introduced a privacy control allowing streamers to prevent their content from being used to train Amazon's generative AI models, according to details first reported by The Verge.

The new "Training for Generative AI" toggle appears in Twitch's Security and Privacy settings. When disabled, the opt-out applies to streams, video-on-demand content, clips, stream chats, and channel images and text. The restriction specifically covers future training of Amazon AI models designed to generate or synthesize text, audio, images, or video.

How the opt-out works

Streamers who enable the privacy control will still have access to AI-powered platform features including automated captions and safety tools. The opt-out is narrowly scoped to exclude content from training generative AI models rather than blocking all AI functionality on the platform.

One limitation: when streamers participate in chat on another creator's channel, that channel owner's privacy preferences determine whether the chat data can be used for AI training. Individual chat participants cannot override the host streamer's settings.

According to Twitch's support documentation, the toggle does not prevent Amazon from using channel content for other purposes outlined in the platform's privacy notice. This includes AI-assisted features for streamer growth and monetization, content recommendations, and community moderation tools like AutoMod.

Why it matters

The opt-out mechanism represents a direct response to growing creator concerns about how their content feeds AI development. As platforms increasingly use user-generated content to train large language models and generative AI systems, the question of consent and compensation has become contentious. By offering granular control over generative AI training specifically, Twitch acknowledges the distinction between operational AI features that benefit the platform and training data extraction that builds Amazon's broader AI capabilities. This approach may set a precedent for how other content platforms handle similar tensions between AI innovation and creator rights.

Default settings unclear

The Verge reported that the toggle appeared enabled by default in at least one user account, though Amazon has not confirmed whether this is the standard configuration for all users. Default opt-in settings have proven controversial at other platforms, as many users never review privacy controls and may unknowingly contribute training data.

The timing of the feature's introduction comes as Amazon expands its generative AI offerings across its product ecosystem, from AWS services to consumer applications. Twitch, which Amazon acquired in 2014, represents a massive repository of video, audio, and text content that could theoretically accelerate AI model development.

Details of the new privacy control were first reported by The Verge.