Twitch Automates AI Training on Creator Content by Default

During a live broadcast featuring Twitch Head of Community Mary Kish and Chief Product Officer Mike Minton, the company admitted that the opt-out mechanism was chosen specifically because an opt-in model would likely see near-zero participation. Minton conceded that he could not confirm whether any streamer data had already been ingested by Amazon’s systems for training purposes prior to the announcement.

Twitch framed the policy update as a new privacy setting, though critics argue the default inclusion ignores the ethical concerns surrounding non-consensual data scraping. While Kish pointed to similar practices at companies like Meta, the move highlights a growing tension between platforms seeking vast datasets and creators who rely on their own likeness and voice as primary intellectual property. Users wishing to exclude their content must navigate to their channel settings, select the security and privacy tab, and manually toggle off the generative AI training option.