YouTube is stepping up its game against AI-generated videos and deepfakes. Starting in May 2026, it will automatically label videos made with photorealistic AI. This change moves beyond the old system, where creators had to disclose AI use voluntarily. Now, YouTube will detect AI content on its own using internal signals and metadata.

The new labels won’t be hidden. They will appear right below the video player on long videos. For Shorts, the label will overlay the video itself. Before, these labels only showed up for sensitive topics like health or politics. Now, every AI-generated video gets a visible marker, no matter the subject.

YouTube also makes some labels permanent. Videos created using YouTube’s own AI tools, like Gemini Omni and Dream Screen, will always carry the AI label. Videos with verified metadata from the Coalition for Content Provenance and Authenticity (C2PA) will too. This metadata tracks a video’s origin and editing history.

The C2PA standard is growing. Big names like Adobe, Microsoft, and OpenAI support it. OpenAI recently partnered with Google to embed invisible SynthID watermarks in AI-generated images and videos. These watermarks are invisible to viewers but can be read by detection systems. YouTube’s detection reads both C2PA data and SynthID signals to spot AI content.

Expanding Deepfake Detection to Everyone

YouTube is not stopping at labeling AI videos. It also expanded its deepfake detection tool, known as Likeness Detection, to all adult users. Until May 2026, only prominent creators, journalists, politicians, and celebrities could use it. Now, anyone 18 or older can enroll to protect their face from unauthorized deepfakes.

The system uses facial recognition. Users verify their identity with a government ID and a short video selfie. Once activated, YouTube scans newly uploaded videos for matches to the user’s face. If it finds a possible deepfake, the user gets a notification. They can review the flagged video and request its removal if it violates their rights.

This tool only detects facial deepfakes. Voice cloning and other biometric fakes are not covered yet. YouTube also does not remove videos automatically. The user must confirm the takedown request, and YouTube’s moderation team reviews each case. Parodies and journalistic uses can stay if they fall under fair use.

How YouTube’s AI Tools Help Creators

YouTube is also adding AI to improve content creation. Its new Gemini Omni neural network is integrated into YouTube Shorts and the YouTube Create app. Gemini Omni helps with editing by automatically adjusting audio, stitching clips, and smoothing transitions. This makes editing easier and faster for creators.

The platform’s AI search feature, called “Ask YouTube,” lets users ask questions instead of typing keywords. It combines Shorts and long videos, then summarizes the results. This feature is in beta for YouTube Premium users in the U.S. It aims to help viewers find better, more relevant content using natural language.

Behind the scenes, YouTube uses many signals to detect AI and deepfake videos. These include metadata tags, sensor noise patterns, encoding fingerprints, and inconsistencies in GPS or timestamps. The system is smart enough to detect attempts to hide AI origins by re-encoding or stripping metadata.

YouTube’s move comes ahead of new European regulations requiring AI transparency. Other platforms like Meta and TikTok also label AI content but rely more on creator disclosure. YouTube’s shift to automated detection shows that relying on creators wasn’t enough.

Still, AI detection isn’t perfect. False positives can happen. That’s why YouTube lets creators contest incorrect labels. For content made with YouTube’s own AI tools or verified metadata, labels are permanent, since those signals are definitive.

Overall, YouTube is working to make the platform more transparent and safer. It wants viewers to know when AI is involved. It aims to protect creators from identity theft and misinformation. And it hopes to keep content creation flowing with new AI tools that help rather than replace human creativity.