Twitch Lets Users Opt Out of AI Training After Backlash Over Amazon’s Default Data Collection
Twitch announced Wednesday that users can now prevent their content from being used to train generative AI models. The platform changed its previously enabled-by-default data-collection option so that streamers and viewers can actively opt out. The feature covers livestreams, clips, images, chat logs and other content within a channel. Users can opt out with one click in their channel settings. Twitch Chief Product Officer Mike Minton said during a livestream on the platform that the company “respects” users’ choices. The most controversial issue was Minton’s explanation of why the option had been enabled by default. “If you change it to opt-in, nobody would agree. That’s the honest truth,” he said. Minton also emphasized that data users choose to retain will not be sold to other companies. Community head Mary Kish told viewers directly: “We don’t expect you to be happy or excited about this. I don’t expect anyone to react positively.” She demonstrated how to disable permission for generative AI training and said that using content to train AI has become an industry standard. Criticism quickly filled Twitch’s livestream chat after the announcement. Many users said, “Nobody would opt in because nobody wants to feed their creativity and content to AI.” Others argued that Twitch had an opportunity to set an industry standard against default data collection but instead chose to compromise. The announcement also publicly confirmed for the first time that Amazon is using Twitch content to train its own AI models. Amazon acquired Twitch in 2014 for nearly $1 billion and has increased its AI investment in recent years as it tries to catch up with other competitors and Meta. It remains unclear when Amazon began collecting the data or how much content has actually been used for model training. Minton acknowledged that he does not know exactly “what Amazon used and what it did not use.” Amid the AI arms race, using platform users’ content to train models has become routine for technology companies. Twitch’s decision to offer an opt-out option is essentially a reactive adjustment made after strong user backlash, rather than a proactive effort to protect creators’ rights. For streamers and viewers, the change at least provides a clear choice. For the platform, however, a default-on system followed by a later remedy has deepened creators’ distrust that their content is being used for free.
The feature covers livestreams, clips, images, chat logs and other content within a channel. Users can opt out with one click in their channel settings. Twitch Chief Product Officer Mike Minton said during a livestream on the platform that the company “respects” users’ choices.
The most controversial issue was Minton’s explanation of why the option had been enabled by default.
“If you change it to opt-in, nobody would agree. That’s the honest truth,” he said.
Minton also emphasized that data users choose to retain will not be sold to other companies. Community head Mary Kish told viewers directly: “We don’t expect you to be happy or excited about this. I don’t expect anyone to react positively.”
She demonstrated how to disable permission for generative AI training and said that using content to train AI has become an industry standard.
Criticism quickly filled Twitch’s livestream chat after the announcement. Many users said, “Nobody would opt in because nobody wants to feed their creativity and content to AI.” Others argued that Twitch had an opportunity to set an industry standard against default data collection but instead chose to compromise.
The announcement also publicly confirmed for the first time that Amazon is using Twitch content to train its own AI models. Amazon acquired Twitch in 2014 for nearly $1 billion and has increased its AI investment in recent years as it tries to catch up with other competitors and Meta.
It remains unclear when Amazon began collecting the data or how much content has actually been used for model training. Minton acknowledged that he does not know exactly “what Amazon used and what it did not use.”
Amid the AI arms race, using platform users’ content to train models has become routine for technology companies. Twitch’s decision to offer an opt-out option is essentially a reactive adjustment made after strong user backlash, rather than a proactive effort to protect creators’ rights.
For streamers and viewers, the change at least provides a clear choice. For the platform, however, a default-on system followed by a later remedy has deepened creators’ distrust that their content is being used for free.