Is that YouTube video clip you’re watching real or created by AI?
YouTube wants to make it easy for viewers to recognize that content on its platform is generated by AI. In 2024, content will be labeled if the creator reveals that they used AI tools. YouTube is now making AI-generated content labels more visible to viewers. It will also start automatically applying labels when it detects that a video contains “highly photorealistic use of AI.”
“We’ve always heard from our community that they value transparency when it comes to generated AI content,” YouTube said in a blog post announcing Wednesday’s update. “These changes are designed to balance transparency and creator control.”
YouTube guidelines will continue to require creators to manually disclose any use of realistic AI. But starting this week, the company will also roll out a new internal system to help identify AI-generated content. “Going forward, we will automatically apply labels when our systems detect significant use of photorealistic AI, even if creators do not explicitly state whether they used AI,” YouTube said in a statement.
YouTube creators who believe their content was incorrectly flagged as AI-generated can change their publishing status using YouTube Studio tools. However, YouTube says that AI labels are “persistent” in some cases, such as content created using YouTube’s own AI tools (such as Veo and Dream Screen) or content that includes C2PA metadata (based on the Coalition’s standards for content provenance and authenticity) that indicates it was generated entirely by AI.
Additionally, YouTube will move the Photorealistic and Meaningful AI Modified or AI-Generated Content disclosure label to a more prominent position. Previously, YouTube labeled AI content in the video’s extended description. Now, for long-form videos, the AI label will appear directly below the video player and above the description. For YouTube Shorts, labels appear as an overlay on the video itself.
YouTube has released an image showing where the new labels will appear.
“The goal here is to get context at a glance. If something looks real, but it’s created by AI, viewers can immediately tell,” Rene Ritchie, head of editorial and creator liaison at YouTube, said in a video about the changes. He added that the AI label alone “does not affect how our videos are recommended or whether they are monetized. This is purely about providing viewers with the right information at the right time.”
Meanwhile, content that YouTube determines is “unrealistic, animated, or slightly altered” (but not entirely AI-generated) will continue to have a disclosure in the detailed description section.
The update comes after YouTube expanded its similarity detection program to all creators (18 and older) earlier this month. It’s designed to help users “discover and control how AI is represented on YouTube.” For creators enrolled in the program, YouTube’s systems identify videos where the creator’s facial likeness may be altered or used synthetically. You can then request removal of “inappropriate content that uses your likeness directly in YouTube Studio.”
