Brief IA

YouTube Automates Tagging of AI-Generated Videos

🤖 Models & LLM·Tom Levy·

YouTube Automates Tagging of AI-Generated Videos

YouTube Automates Tagging of AI-Generated Videos
Key Takeaways
1YouTube will automatically label videos using photorealistic AIs, thereby enhancing control over content.
2AI labels will be more visible, appearing directly below the video player for better identification.
3Creators can correct the labels, but cannot remove them if YouTube's AI tools have been used.
💡Why it mattersThis initiative aims to strengthen transparency and user trust in the face of the proliferation of AI-generated content.
Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisis ton rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

📄
Full Analysis

YouTube has decided to take the lead in transparency on its platform by automating the labeling of videos generated by artificial intelligence. Announced on Wednesday, this measure means that YouTube will no longer rely solely on creators to report the use of AI in their videos. From now on, its internal systems will automatically apply labels when significant photorealistic AIs are detected.

This initiative comes with improved visibility for AI labels, which will now be easier to spot on long videos as well as on YouTube Shorts. For over two years, YouTube has been using these labels after updating its policies regarding AI. Creators were already required to disclose the use of AI in their videos, unless the content was clearly fictional or animated, such as a unicorn dancing in an imaginary world.

Although the labeling policy has not changed, YouTube is committed to taking a more active role in content monitoring. This decision follows Google's launch of Gemini Omni, a new family of multimodal AI models, at the Google I/O conference. These models are capable of producing high-quality videos with a deep understanding of various domains.

Starting in May, YouTube will use new internal signals to identify and label AI-generated content. Creators must still disclose their use of AI, but in cases of omission, YouTube will step in to label the video.

Creators have the option to update the disclosure status of their videos, but they will not be able to remove labels if the content was created using YouTube's AI tools, such as Veo or Dream Screen. Videos containing C2PA metadata, indicating full AI creation, will also receive permanent labels. OpenAI, Nvidia, Kakao, and Eleven Labs have recently adopted this C2PA standard.

This automatic AI detection feature comes shortly after YouTube expanded its deepfake detection capabilities. Now, any adult can search for face matches on YouTube, a feature initially tested with celebrities and public figures.

YouTube is also committed to making its AI labels more consistent and prominent. Previously, they appeared in the extended description, except for videos on sensitive topics like health or current events, where they appeared directly on the video. Now, labels will be visible under the video player for long videos and overlaid on YouTube Shorts.

For slightly modified or unrealistic AI videos, the label will only appear in the extended description. It is important to note that these labels will not affect the recommendation or monetization of the videos.

In parallel with this increased monitoring, YouTube continues to invest in AI to develop features such as its interactive search engine Ask YouTube, a playlist generator for YouTube Music, AI video summaries, and other generative AI creation tools.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.