YouTube to Auto-Detect and Label Realistic AI Videos, Introduces Custom Feeds

As artificial intelligence continues to reshape digital content, YouTube is rolling out a major update aimed at improving transparency and user control. The platform is introducing clearer AI-generated content labels, automated detection systems, and a new personalised feed feature that allows viewers to tailor recommendations based on their preferences, moods, and prompts.

YouTube Strengthens AI Transparency Measures

With AI-generated videos becoming increasingly realistic and widespread, YouTube is taking a proactive approach to ensure users understand what they are watching. While creators have been required to disclose AI-generated content since 2024, the platform is now simplifying the process and making disclosures far more visible.

The company’s latest update focuses on improving clarity for viewers while maintaining flexibility for creators who use AI responsibly. By enhancing label visibility and introducing automatic detection systems, YouTube aims to bridge the gap between content innovation and viewer trust.

More Prominent AI Labels Across the Platform

Clearer Placement for Long-Form Videos

YouTube is redesigning how AI disclosures appear on standard videos. Moving forward, labels indicating AI-generated or AI-edited content will be placed directly below the video player. Additionally, they will appear above the description section, ensuring viewers notice them without needing to expand additional information.

AI Labels for Shorts

For YouTube Shorts, the disclosure system will be even more direct. AI labels will appear as overlays on the video itself, making them immediately visible during playback. This ensures that short-form viewers are equally informed without interrupting their viewing experience.

Exceptions for Minor or Unrealistic Edits

Not all AI-assisted content will receive prominent labels. Videos that include minor AI edits, stylised effects, or unrealistic animations will continue to display disclosures within the expanded description rather than in the main viewing area.

Automatic AI Detection System Launching in 2026

How the Detection System Works

Starting May 2026, YouTube will begin deploying internal AI tools capable of identifying realistic AI-generated content automatically. This marks a significant shift from relying solely on creators to disclose such content.

If a creator fails to label their video appropriately, YouTube’s system will analyse visual elements and apply a disclosure label if substantial AI-generated material is detected.

Creator Control and Appeals

Despite automation, creators will still retain control over their content. If a video is incorrectly flagged, they can update disclosure settings via YouTube Studio. This ensures a balance between platform oversight and creator autonomy.

Permanent Label Scenarios

In certain cases, AI labels will remain permanently attached to videos. These include:

• Content created using YouTube’s own AI tools such as Veo or Dream Screen
• Videos embedded with C2PA metadata confirming full AI generation

YouTube has clarified that these labels are strictly informational and will not impact a video’s monetisation or recommendation performance.

Transparency Without Penalty

A key highlight of YouTube’s update is its emphasis on neutrality. The platform has made it clear that AI labels are not meant to penalise creators. Instead, they are designed to provide viewers with context about the nature of the content.

As AI tools become more sophisticated, distinguishing between real and synthetic media is becoming increasingly difficult. YouTube’s approach focuses on informed viewing rather than restricting creativity.

Introduction of Personalised Custom Feeds

A New Way to Discover Content

Alongside AI transparency updates, YouTube is introducing a new feature that allows users to generate customised recommendation feeds using prompts. This feature gives viewers more direct control over what they see, moving beyond the traditional algorithm-driven suggestions.

Users can simply describe what they want to watch, and YouTube will create a continuously updating feed tailored to that request.

Example Prompts

YouTube has shared examples to demonstrate the feature’s flexibility:

• “Show me something different from my usual content”
• “Help me relax with short guided meditations under 10 minutes”

These prompts allow users to shape their content experience based on mood, curiosity, or specific interests.

How the Custom Feed Feature Works

Using the feature is simple and intuitive:

• Tap on the “Your custom feed” option on the Home page
• Enter a prompt describing desired content
• Choose from suggested prompts if needed
• Edit prompts anytime to refresh recommendations

Once created, these feeds can be pinned as chips at the top of the Home screen for easy access, making personalised discovery more convenient than ever.

Availability and Requirements

Currently, the custom feed feature is rolling out in English for signed-in users in the United States across mobile apps and desktop platforms. To access it, users must keep their watch history and search history enabled.

YouTube has also incorporated a feedback mechanism, allowing users to report mismatched recommendations and improve the system over time.

The Bigger Picture: AI and the Future of Content

YouTube’s latest updates highlight a broader industry trend toward responsible AI usage. As synthetic media becomes more advanced, platforms are under increasing pressure to ensure transparency while fostering innovation.

By combining visible disclosures, automated detection, and enhanced personalisation tools, YouTube is positioning itself at the forefront of this transition.

The move not only empowers viewers with better information but also gives creators the tools to experiment with AI while maintaining trust and credibility.

Conclusion

YouTube’s new AI labelling and custom feed features mark a significant step forward in the evolution of digital content platforms. By prioritising transparency and user control, the company is addressing growing concerns around AI-generated media while enhancing the overall viewing experience.

As these features roll out globally, they are expected to redefine how audiences interact with content in an increasingly AI-driven ecosystem.