YouTube setting new guidelines for videos that contain AI-generated content

Artificial intelligence is rapidly transforming how content is being made. AI models can now write text, generate images, produce videos, and even create synthetic voices that sound shockingly real. As this tech only becomes more and more accessible, companies are starting to grapple with the ethical implications and concerns surrounding transparency. YouTube, the world’s largest video platform, is taking steps to address these concerns with its newly-unveiled rules for labeling AI-generated content.

YouTube’s New AI disclosure rules

Beginning today, YouTube creators will be required to identify when their videos contain AI-generated or otherwise manipulated content that appears realistic. This requirement falls broadly under the umbrella of “altered or synthetic media,” which includes the use of generative AI. Under YouTube’s new guidelines, creators need to label their videos accordingly if they include elements like:

• A synthetic version of a real person’s voice as narration
• Replaced or altered faces
• Manipulated footage of real events or locations
• Prominent Labeling for Sensitive Topics

YouTube will also take additional steps for sensitive topics such as news, politics, health, or finance. Videos in these categories will have a more noticeable label attached directly within the video player. YouTube wants to strike a balance between transparency and practicality. Creators won’t need to disclose AI use in the following scenarios:

• AI tools assisting with scriptwriting or brainstorming
• Automatically generated captions
• “Clearly unrealistic content” (e.g., fantastical animations)
• Minor special effects and standard color or lighting adjustments

Labeling AI-generated content creates much-needed transparency for viewers, ensuring they’re informed of what they’re watching. This is especially important as deepfakes and other synthetic media become more sophisticated and potentially misleading. YouTube’s rules should promote ethical content creation and will help build audience trust. We’re going to need it in the days ahead, for sure.

Penalties for not playing along

Initially, YouTube says it aims to educate creators, not punish them. However, those who repeatedly fail to label AI content properly may face penalties. Additionally, the platform is working on a new takedown procedure for harmful synthetic content, such as realistic deepfakes of identifiable people. These guidelines are likely to evolve over time, but they can hopefully set a precedent for how platforms can handle the evolving and complex world of AI-generated content.