Twitch users are pushing back against Amazon after the livestreaming platform introduced a new setting that allows creators to opt out of having their broadcasts used to train generative AI models.
The controversy began after Twitch added a “Training for Generative AI” option to its privacy settings. The biggest concern is that the feature is opt-out rather than opt-in. This means creators must manually disable it if they do not want their livestreams used as training data.
According to Twitch, information gathered from broadcasts could be used to train future AI models capable of generating or synthesizing text, audio, images and video.
The announcement has frustrated both streamers and viewers. Creators are particularly concerned about their voices, appearances, gameplay and conversations becoming part of AI training datasets without explicit permission.
Twitch Chief Product Officer Mike Minton addressed the backlash during a livestream and admitted that making the feature opt-in would likely result in very few creators participating.
For some streamers, the issue goes beyond their own content. Livestreams often include guests, musicians, voice actors, and developers who may have strong objections to generative AI. Creators are now questioning whether featuring those people could unintentionally contribute their work to AI training.
Some streamers have already begun adding “AIOptedOut” and “NoAI” tags to their broadcasts.
The controversy reflects a much bigger debate across the technology industry: who owns the data created online, and should companies be allowed to use it to build AI systems without asking first?
