Twitch Users Can Now Choose to Stop AI Training on Their Content

Twitch has officially introduced a new way for users to control how their content is used by Amazon. As of today, streamers can choose to opt out of having their videos, clips, and chat messages used to train Amazon’s generative AI models. Generative AI is a type of technology that can create new text, images, audio, or video based on the data it has learned from. This update gives creators more control over their personal work, which has been used by the company for AI development for at least two years.

According to Ars Technica, this change comes after a long period where users were automatically included in the training process. In 2024, Mike Minton, who was the chief monetization officer at Twitch at the time, confirmed that the company was using platform content to help build these models. He stated that the company was working within the rules of user trust and privacy laws, but many users were unaware that their streams were being used in this way until now.

To stop their content from being used, users must visit their security and privacy settings on the Twitch website. The Verge reports that the new "Training for Generative AI" toggle was turned on by default for many users. By turning this setting off, a creator ensures that their future streams, recorded videos, and chat messages will not be used to teach Amazon’s AI how to generate new content. However, it is important to note that if a user participates in someone else’s stream, the rules of that channel owner apply. This means the streamer’s own settings will decide if the chat messages from viewers can be used for training.

It is important to understand what this change does and does not do. Twitch clarified that turning off this specific AI training toggle does not mean the company will stop using all AI tools on the platform. Other "AI-supported" features will continue to function normally. These include tools that help with community safety, such as AutoMod, which automatically filters out bad language in chat. Other features that help streamers grow their channels, like real-time help for sponsorship campaigns and systems that suggest new channels to viewers, will also remain active. The company says these tools are meant to benefit the community by making the platform easier to use and more secure.

This move toward more user control comes at a time when many people are worried about how large technology companies use personal data. There is a growing public debate about whether companies should be allowed to use content created by individuals to build powerful AI systems without asking for permission first. For example, recent reports have shown that other companies, such as Meta, have faced significant public backlash regarding how their new smart glasses record information and how that data might be handled.

For many Twitch creators, the ability to opt out is a welcome change. It allows them to decide for themselves if they want to contribute to the development of new technology. While the company maintains that using this data helps improve services like automatic captions, the new setting ensures that the choice remains in the hands of the user. As AI technology continues to grow and change, the question of who owns the data used to train these systems will likely remain an important topic for both creators and technology companies for years to come.