As AI video technology advances, YouTube is shifting towards a more proactive approach in labelling AI-generated content. The platform recently announced that it will now automatically apply labels to videos when its systems detect the use of “significant photorealistic AI,” rather than relying solely on creators to disclose this information.
Starting in May, YouTube will enhance the visibility of these AI labels, making them easier to find on long-form videos and YouTube Shorts. For over two years, YouTube has utilised AI labels, following an update to its policies and the introduction of a tool in Creator Studio that required creators to announce when their videos contained AI content that might be mistaken for real people, places, or events. However, videos featuring purely animated or imaginative elements, such as a unicorn roaming in a whimsical setting, do not require these labels.
While YouTube states that its AI labelling policy remains unchanged, the company will take a more active stance in monitoring the content shared on its platform. This decision follows Google’s recent introduction of Gemini Omni, a new suite of multimodal AI models capable of producing high-quality videos that incorporate knowledge of physics, history, culture, and science.
In its latest initiative, YouTube will deploy new internal signals to identify AI-generated content. Creators are still encouraged to disclose their use of AI, but if they fail to do so, YouTube will label the video on their behalf. Creators who believe their content has been misclassified can adjust the disclosure status, but they can’t remove labels for videos created with YouTube’s own AI tools, such as Veo or Dream Screen.
Additionally, labels will be permanently attached to videos that contain C2PA metadata indicating that AI entirely generated them. Recently, OpenAI committed to the C2PA standard, joining other companies like Nvidia, Kakao, and Eleven Labs.
This new automatic AI detection feature arrives shortly after YouTube expanded its deepfake detection capabilities, which now allow all adults to search for face matches in videos, following initial trials with celebrities, public figures, and notable creators.
YouTube plans to standardise and enhance the prominence of AI labels. Previously, these labels appeared in the expanded description, unless the video covered sensitive topics like health or news, in which case they were featured more prominently on the video itself. Now, labels will be displayed directly beneath the video player for long-form content and will appear over YouTube Shorts.
This relocation aims to ensure that viewers can easily spot photorealistic, AI-altered, or AI-generated content. In contrast, for content that is only minimally altered, animated, or otherwise unrealistic, the AI label will still appear in the expanded description.
Importantly, YouTube has clarified that the presence of AI labels will not affect a video’s recommendations or its monetisation potential.
In addition to its initiatives for regulating AI content, the company continues to invest in AI technologies for various functionalities, including its interactive search feature, Ask YouTube, a playlist generator for YouTube Music, AI-driven video summaries, and other generative AI tools.
