Amazon’s Twitch Enables AI Training by Default, Sparking Creator Backlash

Twitch, the live-streaming platform owned by Amazon, has drawn criticism from content creators after it emerged that a setting permitting the use of channel content to train Amazon’s generative AI models is enabled by default. Users must manually disable the option if they wish to prevent their material from being used for this purpose.

Default activation sparks debate over consent

According to a report published by Engadget, Twitch introduced a new setting in its user interface that allows channel owners to opt out of having their content—including live broadcasts, archived videos, clips, chat logs, text, and images—used to train generative AI models developed by Amazon. The controversy centers on the fact that this setting is active upon account creation, meaning users are automatically enrolled unless they take deliberate action to turn it off.

Opt-out does not block all AI functions

Twitch clarifies that disabling generative AI training does not halt all artificial intelligence or machine learning applications on the platform. Features such as automated translation tools, content moderation systems, and community safety mechanisms operate independently of this setting. A further complication arises with chat: messages posted by viewers may still be subject to the AI training policy based on the settings configured by the channel owner where those messages appear.

Raising fundamental questions about user control

The implementation renews longstanding concerns about data sovereignty and informed consent. It intensifies an existing debate: should training AI models on user-generated content require explicit, affirmative permission—or should it operate under a default-on model, placing the burden on individuals to locate and deactivate the feature?