Twitch has confirmed that Amazon is utilizing user livestreams to train its artificial intelligence models, prompting significant backlash from creators who are now required to manually opt out.

Key Points

• Twitch users are automatically opted into AI training, requiring them to manually disable the setting within their privacy dashboard.
• The announcement triggered widespread criticism from high-profile streamers and community members on social media platforms like X.
• Twitch executives Mary Kish and Mike Minton confirmed the training practices during a livestream but stated they lack oversight of Amazon’s specific data usage.
• Twitch currently uses AI for internal tools like captions and content classification, though it does not currently employ generative AI features.
• Amazon, which acquired Twitch in 2014, is aggressively scaling its AI infrastructure to compete with industry leaders like OpenAI and Anthropic.

Why it Matters

This development highlights growing tensions between content creators and major technology companies regarding the use of user-generated data for training large-scale AI models. As Amazon accelerates its efforts to compete in the frontier model market, the controversy underscores the increasing demand for transparency and user control over digital intellectual property.