If you stream on Twitch and haven't changed any settings recently — your content is already likely participating in training Amazon's generative AI. The platform didn't ask for permission in advance: the option was enabled by default, and you have to disable it yourself.
Practical part: how to opt out
To remove your channel from the program, you need to:
- Go to your Twitch channel settings;
- Open the Security and Privacy section;
- Find the Training for generative AI parameter and disable it.
The system shows no notifications, confirmations, or reminders — just a toggle switch that you have to find yourself.
Why this isn't a minor issue
A streamer's content is not just video. It's a voice, manner of speaking, reactions, gaming style, sometimes even a face on camera from years of daily broadcasts. For a generative AI model, this is ready-made data for training speech recognition, voice synthesis, or even reproduction of behavioral patterns.
When asked directly whether Amazon had used content from creators to train models before the new rule was introduced, Twitch representatives gave no clear answer. This is itself a telling moment: the trust issue here goes beyond one settings function.
Platform creators fear that thousands of hours of broadcasts will become material for AI without conscious consent — and formally they have a point: consent here exists only in the form of silent refusal to refuse.
Ukrainian context
Twitch holds about 66% of the Ukrainian-language streaming market, and esports streams from Ukraine last year collected over 13 million viewing hours. For thousands of Ukrainian creators who earn from donations and advertising specifically on this platform, changing the rules of the game without explicit notification is a question not only of privacy, but also of control over their own work product.
The question that remains open: will silent consent become the norm for other platforms if the Twitch community's response doesn't lead to a change in default policy?
