AI Pulse by Inblix

Twitch quietly flips on AI training toggle — here's how to opt out

The Verge AI · Aug 12, 2026 · 2 min read · Read original article →

Curated by the Inblix editorial team


Featured image for article: Twitch quietly flips on AI training toggle — here's how to opt out

Twitch has added a new opt-out toggle that lets streamers block Amazon from using their content to train generative AI models. The setting, found under Security and Privacy, covers a lot of ground: streams, VODs, clips, chat messages, and even pictures and text posted on your channel. Amazon’s description says flipping it off prevents your content from being used in “future training” of models that generate text, audio, images, or video.

Here’s the catch that’s already raising eyebrows. Jay Peters at The Verge checked his own settings and found the toggle was switched on by default. That means unless you actively go in and change it, Amazon has permission to scrape your broadcasts and community interactions for AI training. Twitch hasn’t clarified whether that’s the universal default or an account-specific rollout quirk. The company told The Verge it would have more information soon — but as of now, streamers should assume they’re opted in until proven otherwise.

Opting out doesn’t break everything. Twitch says AI-supported features like AutoMod, recommendations, and real-time sponsorship assistance will keep working because those fall under the broader privacy notice rather than generative AI training. But there’s a wrinkle for chatters: if you’re typing in someone else’s stream, their opt-out preference — not yours — controls whether your messages can be used for training. That’s a notable shift in how platform-wide data rights work, and it means lurkers and regulars in large channels have essentially no individual control over their chat history.

This move lands in the middle of a broader reckoning over how platforms treat user content in the AI era. Reddit, Stack Overflow, and WordPress have all faced backlash for either selling or quietly allowing AI training on user-generated material. What makes Twitch’s approach interesting is the transparency — a visible toggle is more than many platforms offer. But default-on consent is the same playbook Meta and Google have used, and it places the burden on millions of streamers to discover a setting buried in a privacy tab. The real test will be whether Twitch ever changes that default, or whether most creators simply never find the switch.

💡 Key Takeaways

  1. The AI training opt-out toggle was found switched on by default, meaning most streamers are likely opted in without realizing it.
  2. Opting out blocks streams, VODs, clips, and channel images from training generative AI models, but AI features like AutoMod and recommendations still function.
  3. Chat messages posted in another streamer's channel are governed by that streamer's opt-out choice, leaving individual chatters without direct control.
  4. Twitch's default-on approach mirrors Meta and Google's playbook, putting the burden on users to discover and flip a buried privacy setting.

Keep reading: See related articles below for more coverage on this topic.

Get smarter about AI

The sharpest AI news, curated daily. Delivered free to your inbox.

Learn more

Glossary terms

← Back to all articles