Twitch has launched a new privacy control allowing creators to prevent their content from being used for training Amazon’s artificial intelligence systems. This opt-out option, found under the "Generative AI Training" toggle in the Security and Privacy section of Twitch’s account settings, responds to growing unease among creators about how their streams, videos, and posts are utilized.
Despite the new option, disabling this setting does not stop all AI uses of content on Twitch. The platform and Amazon retain rights to apply creator material for other purposes, such as AI-powered sponsorship recommendations, content discovery algorithms, and moderation tools like AutoMod, as outlined in Twitch’s Privacy Notice. This partial restriction has fueled backlash, with thousands of creators expressing frustration that their content was previously included by default without explicit disclosure.
The controversy highlights ongoing questions about consent and transparency in AI training practices. Twitch’s Terms of Service, revised earlier this year, grant the company broad rights to reproduce, modify, distribute, and create derivative works from user content. However, these terms did not specifically specify the use of creator content for generative AI model training until this latest update. Similar concerns have emerged with other major tech firms, including Meta and Google, which have faced scrutiny over incorporating user-generated material into AI tools.
From a practical standpoint, creators who wish to protect their work from AI training can adjust the new setting via the Twitch website or mobile app, but public content remains vulnerable to scraping or analysis by other entities. The move signals a broader shift in digital platforms gradually seeking explicit consent for AI use, albeit often after the technology has already been deployed extensively.
Twitch’s product leadership has acknowledged that the new control might face criticism but considered it necessary to maintain participation across the platform. As Twitch explained, most publicly available content feeds AI model training across the industry, complicating efforts to fully shield individual creators’ work.