Twitch streamers can now opt out from training Amazon’s AI

Twitch has introduced a new user setting that enables content creators to prevent their digital media from being utilized to train Amazon's generative artificial intelligence systems.

By adjusting this control, streamers can ensure that their live broadcasts, recorded videos, clips, chat logs, and channel images will not be incorporated into future datasets used to train machine learning models designed to generate synthetic text, audio, images, or video. Despite opting out, creators will still have access to other AI-driven platform utilities that are not generative in nature, such as automated closed captions, community recommendation engines, and standard moderation assistance tools.

The newly integrated setting, labeled "Training for Generative AI," is located within the platform's security and privacy dashboard. Early observations indicate that the feature is enabled by default, requiring creators to take active steps to change their preferences if they wish to restrict their content. Twitch also clarified that the data usage policy for interactive features remains dependent on the channel host; for instance, if a viewer participates in a chat on another stream, that specific broadcaster's preferences will determine whether the chat transcripts are fed into Amazon's training models. This update reflects broader ongoing discussions within the technology sector regarding data ownership, creator consent, and the ethical collection of training materials for modern AI software.