Twitch has confirmed it uses creator streams to train artificial intelligence models, a practice now under scrutiny from its content creator community. According to Engadget, the platform did not make this data usage opt-in, meaning streams are included by default unless creators take action to opt out. This has raised concerns among streamers about how their content is being used to develop AI systems without explicit consent.

To stop Twitch from training AI on your streams, creators must navigate to their account settings and disable the relevant data sharing option. The setting is located under Privacy and Safety, where users can toggle off "Allow Twitch to use your content to improve its services." Disabling this prevents your streams from being included in AI training datasets moving forward.

The lack of an opt-in approach has sparked debate about creator rights and platform transparency. Many streamers argue that using their live content—often the product of significant time and effort—to train AI should require clear, affirmative permission. Critics note that opt-out models place the burden on creators to monitor and adjust settings they may not even know exist.

Twitch has not disclosed specific details about how the AI is being used, what models are being trained, or whether the data includes audio, video, chat, or metadata. However, the company maintains that the data helps improve platform features such as content recommendations, moderation tools, and accessibility features like automated captions.

For creators concerned about intellectual property and data ownership, reviewing and adjusting privacy settings is now a necessary step. As AI integration grows across streaming platforms, understanding how your content is used— and having control over that use— remains a critical part of managing a sustainable creator business. Staying informed about policy changes and platform practices is essential for protecting both creative work and audience trust.