Twitch makes it practically impossible to prevent AI scraping

2 Min Read

Amazon-owned streaming platform Twitch, used by many to broadcast and promote video games, is facing criticism for harnessing user data to train the generative AI models of its parent company.

In a statement on X, the company confirmed users will need to manually opt-out to prevent their content being used to train generative AI models across Amazon. In a follow-up stream (a clip of which is being widely circulated online), Twitch chief product officer Mike Minton attempted to answer the glaring question: why is it not opt-in?

“Now why is it not opt-in? That’s what everybody is here spamming in chat. I get it. Let me opt-in versus making me opt-out. Well, there’s an honest answer and I think most of you probably can appreciate this: if it was opt-in nobody would opt-in. That’s honestly the answer. It’s going to be on by default, and like I said, almost every content service in the world is on by default,” said Minton.

Related:Rockstar wants you to pay Netflix to watch the next GTA VI trailer before it’s released to the masses

More information was shared by Twitch in an FAQ, with the company noting that even if a user chooses to opt-out, their content might still be used for training purposes if they appear on someone else’s stream.

“Your stream and the stream’s chat, your VODs, your Clips, Highlights, and any text or images on your Channel may be used. If you chat on someone else’s stream, their opt-out preferences govern if that chat can be used for training,” reads the explainer.

In short, then, Twitch has not provided a surefire way for users to prevent their content being scraped for generative AI training—aside from, presumably, ditching the platform entirely.

Indeed, as reiterated on that same FAQ, if someone does choose to opt-out, that decision will only extend to streams, VODs, clips, stream chats, and pictures and text on their own channel.

“Opting-out of training generative AI content models does not opt you out of all AI or machine learning uses at Twitch. For example, Twitch will still run AutoMod, which helps keep our community safe, but does not retain your data and use it to produce content,” the FAQ adds. “Twitch may also still use your data to run AI-supported features, like captions, which automatically generate captions for your Clips but does not retain your stream or your clips to train a model that can generate new content.”

The move also raises another question: how can video game studios who don’t agree with generative AI usage prevent their work being streamed on the platform and subsequently used to train Amazon’s models—without, perhaps, hip-firing DMCA takedown requests?

Related:EA announces advertising platform to launch ads ‘directly into gameplay’

As noted by GamesIndustry.biz, Twitch captured over 60 percent of global video game live-streaming audience in 2024. Viewers watched over 15.6 billion hours of game content on the platform during that period.

Game Developer has reached out to Twitch for more information.

About the Author