YouTube Unveils New AI Training Controls for Creators (Third-Party Training)
YouTube announced a major update on Monday that gives creators more control over how third-party companies can use their content to train AI models. This new feature will allow creators to choose whether they want to permit specific companies to train their AI models using their videos.
Starting today, YouTube creators will find a new setting in the YouTube Studio dashboard, allowing them to easily manage which third-party AI companies can use their content for training purposes. This new feature offers creators more choice and transparency, allowing them to select from a list of 18 AI companies, or an “All third-party companies” setting, granting any company permission to train on their data.
Here’s the initial list of 18 AI companies that were approved to train: some of the biggest names in the field are included, like
- AI21 Labs
- Adobe
- Amazon
- Anthropic
- Apple
- ByteDance
- Cohere
- IBM
- Meta
- Microsoft
- Nvidia
- OpenAI
- Perplexity
- Pika Labs
- Runway
- Stability AI
- xAI
These companies are at the forefront of developing generative AI models, and YouTube has identified them as sensible partners for content creators. However, creators are not limited to the listed companies and can choose to allow all third parties to train on their videos if they prefer.
Who Can Access These Settings?
The creator will have to have a YouTube Studio Content Manager that holds an administrator role to apply this feature. Once implemented, the creators can readily alter their AI training at any given time through Channel Settings on YouTube Studio.
This is in response to concerns from creators who have raised issues about companies using their content without consent or compensation to train AI models. YouTube’s new feature looks to address these concerns through a clear mechanism for giving or denying permission for training AI on their content.
Default Settings and Google’s Role
By default, the setting will not allow third-party companies to train on the content of a creator. It will make it very evident to companies that have possibly used creator material without consent that they are not permitted to do so unless there is explicit consent. But YouTube will still use the creator content to train its AI models based on its current agreements.
While the new environment does help control third-party access, it does not amend YouTube’s Terms of Service, which still prohibits third parties from scraping or otherwise using creator content without authorization. YouTube also explains that this is only the beginning in granting creators more control over how their content is accessed and used. In the future, the platform may be allowed to download videos directly for training by authorized companies.
Future Upgrades to Creator Control
This is part of YouTube’s broader efforts to protect creators’ rights in the age of AI. Earlier this year, YouTube unveiled AI detection tools designed to prevent the unauthorized use of creators’ likenesses, including faces and voices, in AI-generated videos. These tools will work alongside the platform’s existing Content ID system, which helps detect and manage copyright violations.
YouTube is set to inform creators worldwide of this new feature by showing banners in YouTube Studio in both desktop and mobile applications in the coming days.
Key Takeaways:
YouTube is introducing a new setting that will allow creators to control which third-party AI companies can use their content for training.
Creators will be able to opt-in to allow AI companies such as OpenAI, Apple, Adobe, Microsoft, and others to train AI models on their videos.
The default setting will prevent third parties from using creator content unless they are explicitly permitted to do so.
YouTube will continue to use content for its own AI training under existing agreements, but the feature represents a step toward greater control for creators.
