Twitch Streamers Must Opt Out of AI Training for Amazon Content

Twitch, the popular live-streaming platform owned by Amazon, has implemented a new policy allowing its parent company to utilize creators' content for training generative artificial intelligence models. This decision has ignited significant controversy within the Twitch community, primarily because users are automatically opted into this data usage unless they actively choose to opt out.

For Amazon, the vast archives of video and audio content generated by Twitch streamers represent an invaluable resource, offering potentially thousands of hours of material to refine and enhance its AI capabilities. However, the platform's default opt-in approach has raised concerns among users who fear their content might be used for training Amazon's AI without their explicit consent or knowledge. This is particularly worrisome on Twitch, where many creators broadcast for extended periods, often capturing personal interactions and voice data.

During a live stream on the official Twitch channel, Head of Community Mary Kish and Chief Product Officer Mike Minton addressed nearly 3,000 viewers expressing strong opposition to the new policy. "Why is it not opt-in? That’s what everybody is spamming in chat. I get it. ‘Let me opt in versus making me opt out,’" Minton acknowledged. He offered a candid explanation: "Well, there’s an honest answer… If this was opt-in, nobody would opt in. That’s honestly the answer."

Community Backlash and Amazon's Rationale

Twitch is aware that a significant portion of its creator base is apprehensive about generative AI, especially given that many prominent AI products have been trained on internet data often scraped without explicit permission. The platform's communication strategy for this change also indicated an expectation of negative reactions. Rather than announcing Amazon's intent to train AI on Twitch content, the platform framed the update as adding "a setting that lets you opt out of having your channel content used to train generative AI content models across Amazon." This phrasing caused confusion, with some streamers questioning if their content had already been incorporated into AI models.

When questioned about whether past videos had already been used for training, Minton admitted, "I don’t actually know the answer to that question because I don’t know what Amazon […] has done in terms of model training and what they’ve used and not used." Kish emphasized that Twitch's move is not an isolated incident, pointing to platforms like Meta, which uses public content from Facebook and Instagram to train its AI. For users in the UK, opting out of Meta's training is possible, but for others, only private settings offer a similar protection, which is impractical for creators who rely on public visibility for monetization.

Kish stated that offering an opt-out option demonstrates Twitch's responsiveness to community sentiment. "[This] is a reflection of reacting to this community’s voice that you are not wanting to train Gen AI models, and we want to give you that option," she explained. To disable the feature, users must navigate to their channel settings, not the creator dashboard, find the 'Security and Privacy' tab, and then toggle off the "training for generative AI" option.

The debate highlights a growing tension between platforms leveraging user-generated content for technological advancement and creators' rights to control their digital output. As AI technology continues to evolve, such issues surrounding data ownership and consent are likely to remain at the forefront for online communities and content creators worldwide.