By Interestana AI Editorial — AI-drafted, human-overseen. How we report
Twitch Users Can Now Opt Out of AI Training

Twitch has introduced a new feature that allows its users to opt out of having their content used for training Amazon's generative AI models. This change, announced on October 26, 2023, addresses concerns that have been present for over two years, since a company executive confirmed Amazon's use of Twitch content for AI development. The updated support page on Twitch explicitly states that streamers must actively opt out if they do not want their "streams, VODs, clips, stream chats, and pictures and text on your channel" to be utilized in the future training of models developed by Amazon. These models are designed to generate or synthesize various forms of content, including text, audio, images, and video.
To exercise this new control, users are directed to a specific section within their account settings. The opt-out process requires navigating to www.twitch.tv/settings/security. This move by Twitch signifies a response to growing user privacy concerns regarding the use of their creative output for artificial intelligence development. Previously, the use of Twitch content for AI training was confirmed by a company executive, sparking debate about data ownership and consent. The platform's decision to implement an opt-out mechanism provides streamers with a more direct way to manage how their intellectual property contributes to the advancement of AI technologies.
Amazon's generative AI content models are part of a broader industry trend where large technology companies are leveraging vast datasets, including user-generated content from platforms like Twitch, to build and refine their AI capabilities. These models are crucial for developing a wide range of AI applications, from virtual assistants and content creation tools to more sophisticated analytical systems. The ability for users to control the use of their data in this context is becoming an increasingly important aspect of digital platform governance. The opt-out feature on Twitch offers a concrete step towards greater user agency in the age of AI, allowing creators to decide whether their work will contribute to the training datasets of these powerful AI systems.
The implications of this policy change extend beyond individual streamers. It reflects a broader conversation about the ethical considerations surrounding AI training data and the rights of content creators. As AI technology continues to advance rapidly, the sources of data used to train these models are under increasing scrutiny. Twitch's decision to provide an opt-out option may set a precedent for other platforms that utilize user-generated content for AI development, encouraging more transparent and user-centric data practices across the digital landscape. The specific types of content that can be opted out include live streams, video-on-demand recordings, short clips, chat logs, and any images or text displayed on a user's channel, underscoring the comprehensive nature of the opt-out provision.
Original source — read the full reporting at the publisher:
Read on Ars TechnicaGet the weekly AI digest
AI news + new model releases, weekly. Drafted by our agents, reviewed by humans.