Twitch has introduced a “Training for Generative AI” account setting that lets creators opt out of having eligible channel content used to train Amazon generative-AI models. PC Gamer and Windows Central reported that the default-on design, and comments attributed to Twitch Chief Product Officer Mike Minton about opt...
Twitch has added a “Training for Generative AI” control to its account settings, allowing creators to manage whether content from their channels can be used to train Amazon generative-AI content models.
The setting has drawn criticism because it is enabled by default, according to PC Gamer. Under that approach, creators who want to prevent eligible channel content from being included must find the setting and switch it off themselves.
Windows Central reported that Twitch Chief Product Officer Mike Minton discussed the decision during a livestream. According to the publication, Minton said an opt-in model would attract little participation, arguing that “nobody would opt in.” The reported remark sharpened objections from creators who believe platforms should obtain affirmative permission before using their work in AI training.
Twitch is owned by Amazon, making the policy particularly significant for a platform whose creators produce a mix of gameplay, commentary, livestream performances and other original material.
PC Gamer reported that Twitch introduced a default-enabled setting through which channel content may support training for Amazon’s generative-AI content models. Its report said creators can opt out through account settings.
Twitch’s own Account Settings help page documents the existence of a “Training for Generative AI” control. The company’s documentation confirms that the preference is available within account settings, giving users a place to review their selection.
Windows Central independently reported on Minton’s comments about the default choice. Its account matches the broader reporting that Twitch adopted an opt-out structure rather than asking creators to actively enroll in the program.
The available material does not establish every detail of the policy. It does not specify all categories of eligible content, identify particular Amazon models that may use the material, or explain retention practices. Those questions would require further clarification from Twitch’s current help materials, terms or privacy disclosures.
The dispute is part of a wider argument over consent for AI training. Platforms increasingly hold large collections of user-created material that could be useful for developing AI systems, but creators may view that material as an extension of their work, identity and business.
An opt-out design can be practical for a company seeking broad participation, but it places the responsibility on users to notice the policy and act. That can be difficult for occasional streamers, people who do not regularly review account settings, or creators managing channels through multiple tools.
The issue is therefore not solely whether Twitch offers a choice. It is also whether the company makes that choice clear enough for creators to make an informed decision before their content is included. For streamers, the default setting may affect trust in how the platform handles material they produce for audiences and sponsors.
The key developments will be whether Twitch changes the default selection, publishes more detail on eligible content, or explains how Amazon’s AI systems may use training material. Responses from prominent broadcasters, creator groups and game publishers could add pressure for more disclosure.
Creators who want to assess their position can review the “Training for Generative AI” preference in Twitch account settings and consult Twitch’s latest documentation. Because platform policies can change, users should check the current help page and applicable terms rather than relying only on initial reporting.
Hero image prompt: Editorial digital illustration of a livestream creator seated at a desk before a glowing monitor, with abstract blue and purple data streams branching in two directions: one toward a subtle AI-network motif and one toward a privacy-control symbol. Restrained newsroom visual style, dramatic lighting, no text, no logos, no recognizable user interface.
Hero image alt text: An editorial illustration of a livestream creator facing branching abstract data streams leading toward an AI network and a privacy-control symbol.
What happened Twitch has added a “Training for Generative AI” control to its account settings, allowing creators to manage whether content from their channels can be used to train Amazon generative AI content models.
The setting has drawn criticism because it is enabled by default, according to PC Gamer.
Under that approach, creators who want to prevent eligible channel content from being included must find the setting and switch it off themselves.
Continue reading
