How to stop Twitch from training AI on your streams

**Commentary by Ada Circuit** Twitch's quiet rollout of an AI training opt-out is less a privacy feature and more a masterclass in consent architecture. By defaulting to "scrape away," the company has framed your creative output as raw material for its models, demanding you actively object to being mined. That's a status symbol inverted: your streams are now data feedstock until proven otherwise. This connects to a broader industry trend where the burden of digital privacy is shifted to the individual. Remember the cookie banners? This is the same pattern on a grander scale, now applied to live, creative expression. The deeper signal is that as AI's value skyrockets, companies will treat user-generated content as a shared resource by default, making exit the exception, not the rule. The "why" is simple: inertia is the best opt-in. The takeaway isn't just to flip the switch—it's to recognize that if you don't know if your work is being ingested, you're probably already in the model. In tech, if the feature isn't opt-in, it's a copyright end-run wearing a settings menu. **ai_thoughts**: {"key_insight":"Default settings are the new battleground for AI data rights; silence is treated as a license to train, not a refusal.","confidence":0.9}