Twitch will by default allow creators’ livestream recordings to be used to train generative AI models for its parent company, Amazon. The change has provoked swift backlash from the Twitch community because creators are opted in by default and must manually opt out if they do not want their content used.
For Amazon, Twitch stream recordings represent a valuable dataset: thousands of hours of audio and video that can be used to train generative models. Creators worry that the requirement to opt out manually means many could unknowingly have their streams used in AI training. This is particularly sensitive on a platform where streamers regularly record themselves and their voices for many hours each week.
Company response and community reaction
Mary Kish, Twitch Head of Community, and Mike Minton, Twitch Chief Product Officer, addressed the situation in a live stream on Twitch’s official channel, attended by nearly 3,000 viewers. Many viewers posted anti-AI messages in the chat.
When asked why the setting is not opt-in, Minton said: “If this was opt-in, nobody would opt in. That’s honestly the answer.”
Twitch appears to be aware that much of its streamer community opposes generative AI use, since many prominent generative AI systems are trained on books, images, videos and other internet content collected without creators’ consent.
Twitch framed the announcement in terms of adding an opt-out setting rather than saying Amazon would begin using Twitch users’ content for training. That wording caused confusion for some creators, who asked whether their content had already been used. When one user asked if their videos had already been used for training, Minton replied, “I don’t actually know the answer to that question because I don’t know what Amazon […] has done in terms of model training and what they’ve used and not used.”
Comparison with other platforms and rationale for opt-out
Kish pointed out that Twitch is not unique in using user content for AI training. She cited Meta as an example: Meta uses public content from its platforms to train its own AI models, so public Facebook and Instagram accounts have likely already contributed data to Meta’s training. Kish noted that in the U.K. there is an option to opt out of Meta’s training; where direct opt-out is not available, the only practical alternative is to make an account private, which is not feasible for creators who monetize their profiles.
Kish said the inclusion of an opt-out option on Twitch is “a reflection of reacting to this community’s voice that you are not wanting to train Gen AI models, and we want to give you that option.”
How to opt out on Twitch
To opt out of generative AI training on Twitch, creators can:
- Go to their channel settings (not the Creator Dashboard).
- Select the Security and Privacy tab.
- Scroll to the “training for generative AI” option.
- Toggle it off.
This setting allows creators to indicate they do not want their channel content used to train generative AI models within Amazon.
Open questions remain about whether content was used previously for training and whether previously used data can be withdrawn. Twitch executives acknowledged during the live stream that they do not have full answers to those technical and historical questions. The episode is part of a broader debate over how platforms should handle creators’ content, transparency and consent when it comes to AI training.
