Twitch has introduced a new setting in its account settings that lets streamers opt out of having their content used to train artificial intelligence models developed by Amazon, the company's parent firm. The change follows growing concerns among creators about how their data is being used. The opt-out process is straightforward, requiring users to navigate to the Security and Privacy section of their account settings and toggle off the Generative AI Training option. According to Twitch, disabling this option does not stop the platform and Amazon from using the content for other purposes, such as AI-powered features that help streamers grow and monetize their channels.
The new setting aims to give creators more control over how their content is used, but it has also sparked questions about how Twitch and Amazon have used the data up to now. In a forum dedicated to the topic, over 16,000 creators expressed opposition to the default use of their content for AI training, a practice that became public after the update. The backlash followed a livestream in which Mary Kish, Twitch's head of community, acknowledged the change would provoke a negative reaction. Mike Minton, Twitch's head of product, added that keeping the option enabled by default was necessary, as otherwise 'no one would participate' in the process. He also noted that similar practices are likely occurring across other companies developing AI systems.
Twitch’s Terms of Service, effective since March 2024, have allowed the platform and its sublicensees to use, reproduce, modify, and distribute users’ content. However, they have not previously stated that such materials could be used to train generative AI models. This case highlights a broader challenge for AI developers: the growing shortage of high-quality data for training models. Companies like OpenAI have reached agreements with publishers to use their content, but as demand increases, alternatives are emerging, raising ethical and transparency concerns. Similar cases have been documented involving Meta, Google, and YouTube, where user-generated content is being used for AI training without explicit permission.
Source: wired
