YouTube has taken a significant step in empowering creators by providing them with more control over how third parties use their content for training AI models. This move comes in response to the growing concerns among creators regarding the unauthorized use of their material by companies like Apple, Nvidia, and OpenAI. Starting today, creators and rights holders can flag for YouTube if they permit specific third-party AI companies to train models on their content.
Empowering Creators with AI Training Choices
Accessing the New Feature
Creators with access to the YouTube Studio Content Manager and an administrator role can opt into this new feature. From within the YouTube Studio dashboard, they will see a list of 18 companies they can select as having authorization to train on their videos. These initial companies include AI21 Labs, Adobe, Amazon, Anthropic, Apple, ByteDance, Cohere, IBM, Meta, Microsoft, Nvidia, OpenAI, Perplexity, Pika Labs, Runway, Stability AI, and xAI. Creators also have the option to select a setting that allows "All third-party companies" to train on their data, even if not listed.They can view or change their third-party training settings within their YouTube Channel settings at any time, giving them full control over this aspect of their content.
Addressing the Rise of AI Technology
Following the rise of AI technology and the emergence of AI video like OpenAI's Sora, YouTube creators expressed their dissatisfaction with the unauthorized use of their material. YouTube announced its intention to address this issue in the near future. While the new setting controls access by third parties, Google will still train its own AI models on some YouTube content in accordance with existing agreements. The new setting does not change YouTube's Terms of Service, which prohibits unauthorized access to creator content.This feature is seen as the first step towards making it easier for creators to permit companies to train AI on their videos and potentially receive compensation. In the future, YouTube plans to allow authorized companies to access direct downloads of creators' videos.
Default Setting and Retroactive Impact
With the introduction of this feature, the default setting for all creators is that third parties are not allowed to train on their videos. This makes it clear to companies that have already used creators' content without permission that they did so against the creators' wishes. YouTube is unable to say if the new setting has any retroactive impact on existing third-party AI model training. However, its Terms of Service indicates that third parties cannot access creator content without authorization.This new control over AI training is an important step in protecting creators' rights and ensuring that their content is used appropriately.
Previous Announcements and Detection Tools
In September, YouTube first unveiled its plans to offer creator controls for AI training and also announced new AI detection tools. These tools aim to help creators, artists, musicians, actors, and athletes protect their likenesses, including their faces and voices, from being copied and used in other videos. The detection technology builds upon YouTube's existing Content ID system, which previously focused only on copyright-protected material.Creators globally will be notified of the new feature through banner notifications in YouTube Studio on desktop and mobile over the next few days.
