Twitch has introduced a new setting allowing users to opt out of Amazon’s use of their channel content to train generative AI models. The update comes more than two years after a corporate executive first confirmed that Amazon was leveraging Twitch data for AI development.
The Scope of Data Collection
According to Twitch’s updated support documentation, users must proactively opt out if they wish to prevent their streams, VODs, clips, stream chats, and channel images from being used in future training. This data feeds into Amazon-developed models designed to generate or synthesize text, audio, images, or video.
The platform explains that allowing training can lead to functional improvements. For instance, a creator's audio may be used to refine speech-to-text models, which in turn improves captioning services across both Twitch and the broader Amazon ecosystem.
Default Enrollment and Transparency Concerns
A significant point of contention remains the fact that Twitch users are automatically opted into AI training by default. This practice has likely been in place for years; in 2024, Mike Minton, then Twitch’s chief monetization officer, admitted to the practice, stating it was conducted within the bounds of privacy regulations and user trust.
Despite this, online discussions leading up to the announcement revealed a widespread lack of awareness among creators regarding how their content was being utilized. While some users are comfortable contributing to AI development, others cite concerns over the lack of financial compensation, the potential for AI to cannibalize their business, or general skepticism toward Amazon’s data practices.
