Twitch Adds Default AI Training Setting for Amazon Content Models
Twitch has added a channel-level setting that lets creators opt out of having their content used to train generative AI content models across Amazon. The policy puts creators into the arrangement by default, making a manual change necessary for streamers who do not want their channel material included.
The decision concerns a substantial pool of audio and video. Twitch streams often capture creators’ voices and on-camera activity for many hours each week, while archived and live material can provide Amazon with large volumes of training data for generative AI systems.
Opt-out design prompts a community backlash
Twitch addressed the change during a stream on its official channel featuring Head of Community Mary Kish and Chief Product Officer Mike Minton. A live audience of nearly 3,000 users reacted in chat, with many objecting to the platform’s decision to make participation the default.
Minton acknowledged the main complaint: creators wanted an opt-in choice rather than an opt-out control. He said that, if the programme required opt-in, “nobody would opt in,” presenting that as the reason for the design.
The announcement itself described the policy as the addition of a setting that enables users to opt out. That wording led some streamers to ask whether Amazon had already trained on their videos. Minton said he did not know what Amazon had done in model training or which materials it had used.
How creators can change the setting
Kish said the opt-out control reflected Twitch responding to community members who did not want to train generative AI models. She also noted that other platforms use public user content for AI training, citing Meta’s use of public material from its services.
On Twitch, the control is located in Channel Settings, rather than the Creator Dashboard. Users must open the Security and Privacy tab, scroll to Training for generative AI, and turn the option off.
Business implication
Teams and professional creators using Twitch should review the setting for every relevant channel, establish who owns privacy-control decisions and record the choice, particularly where streams include proprietary discussions, staff voices or reusable video assets.