YouTube AI content policy: monetisation, labels and likeness in 2026
What can be monetised, what must be disclosed and how to protect a face: YouTube’s rules on AI content as of November 2026.
YouTube’s AI content policy does not ban AI. It bars “inauthentic” content, meaning mass-produced or repetitive videos, from earning money, and it requires creators to disclose realistic content that has been altered or generated. Since May 2026 the AI label has been far more visible and YouTube applies it itself when it detects undisclosed photorealistic AI; the same month, likeness detection opened to every creator aged 18 and over.
So there are three rules that fit together: what can be monetised, what must be disclosed and how to protect a face. Here they are as of November 2026, with the official sources and a checklist for brands and creators.
YouTube AI content policy: what can be monetised
On 15 July 2025 YouTube updated the rules of the YouTube Partner Program (YPP), the scheme that lets channels earn from advertising. The repetitive content category was renamed “inauthentic content”: mass-produced or repetitive videos, which are not eligible for monetisation. YouTube has always asked for original, authentic content; the update was about spotting what isn’t.
Rene Ritchie, YouTube’s head of editorial and creator liaison, called it a “minor update” to rules already in place. YouTube’s examples are clear: channels uploading narrative stories with only superficial differences between them, or slideshows with the same narration. Reaction and clip videos can still be monetised if they add significant original commentary or modification.
The target is what the press calls AI slop. When the change was announced, TechCrunch pointed to AI voiceovers laid over recycled photos and clips, fake AI-generated news videos and true-crime series that were entirely synthetic.
AI is not the problem
YouTube says it welcomes creators who use AI to tell better stories. In January 2026 CEO Neal Mohan wrote that in December more than a million channels a day had used the platform’s AI creation tools. In the same letter he promised to reduce low-quality, repetitive AI content by extending the systems already used against spam and clickbait.
A brand channel often doesn’t live on ad revenue, but the logic still applies: ten near-identical videos with the same synthetic voice are exactly the profile those systems look for. We cover this in our piece on AI slop.
When an AI video must be disclosed on YouTube
The rule turns on realism, not on the tool. You must disclose content that makes a real person appear to say or do something they didn’t, alters footage of real events or places, or generates a realistic scene that never happened. According to the documentation reported by PPC Land, you don’t need to disclose beauty filters, colour correction, AI-assisted scripts or rough cuts, or clearly unrealistic content such as animation or fantasy.
| Case | Disclose? |
|---|---|
| Testimonial generated with a real person’s face | Yes |
| Real product in a realistic city never filmed | Yes |
| Edited footage of a real event | Yes |
| Ad in clearly unrealistic animation | No |
| AI-written script, shot for real | No |
| Beauty filter or colour correction | No |
Disclosure happens in YouTube Studio: during upload, you answer the AI use question in the video’s attributes. For videos already published, you can change the answer from the content tab.
AI labels in 2026: where they appear and who adds them
On 27 May 2026 YouTube changed its labels. On long-form videos the label now appears below the player, above the description; on Shorts it is overlaid on the video. This is now the single label format for all photorealistic content that has been meaningfully generated or altered. For unrealistic, animated or lightly edited content, the disclosure stays in the expanded description.
The bigger change is that YouTube no longer waits for the creator. From May it is using internal signals to identify generated content: if a video discloses nothing but the system detects significant photorealistic AI, the label is applied automatically. Creators who think it is a mistake can change the status in YouTube Studio.
In two cases the label cannot be removed: content made with YouTube’s own AI tools, such as Veo or Dream Screen, and content whose C2PA metadata marks it as fully generated. For how that metadata works, see our guide to Content Credentials and C2PA.
A label is not a penalty
YouTube is explicit: a label on its own doesn’t change how a video is recommended or whether it can earn money. The risk lies in not disclosing. According to the rules reported by PPC Land, creators who consistently choose not to disclose may have a label applied for them, content removed or be suspended from the Partner Program. The wording targets repeated behaviour, not a single oversight.
Likeness detection: how to protect faces and talent
Likeness detection is the tool YouTube uses to find altered or synthetic versions of an enrolled person’s face in uploaded videos. It is modelled on Content ID, but it works on faces rather than copyrighted works. It was introduced in December 2024 with the agency CAA and in April 2025 was tested with creators such as MrBeast and Marques Brownlee.
- October 2025: a first group of Partner Program creators.
- March 2026: selected creators, government officials, journalists and political candidates.
- April 2026: agencies and management firms such as CAA, UTA and WME and the celebrities they represent, even without a channel.
- May 2026: every creator aged 18 and over, rolled out gradually.
To enrol, you open YouTube Studio on a computer and go to the content detection section: you scan a QR code, then submit a government ID and a selfie video. Possible matches appear in the same tab, and you can request removal for each one.
Two limits to know. The tool recognises faces, not voices on their own. And a match does not trigger automatic removal: parody, satire and other public-interest uses remain protected. The legal picture is in our piece on voice and likeness rights.
Checklist: staying compliant with AI video on YouTube
- Decide in the brief which scenes are synthetic and realistic. Those are the ones to disclose, and knowing up front saves arguments later.
- Answer the AI question on every upload, Shorts included, and align the answer across brand, agency and any creators involved.
- Give every video its own idea. No runs of clips with the same structure and the same voice, with only the text swapped.
- Keep ad-testing variants off the channel’s schedule. They serve the campaigns, not the audience that follows you.
- Keep C2PA metadata where you have it. It documents provenance and triggers the right label.
- Get releases signed for every real face and voice, and encourage the talent you work with regularly to enrol in likeness detection.
- Remember Europe. The transparency obligations of the AI Act apply on every platform, not just YouTube.
This is general information about platform policies, not legal advice.
The rule behind the rules
All of these policies say the same thing: YouTube accepts creators who use AI to make a better video and cuts off revenue for those who use it to make a thousand identical ones. Disclosing doesn’t cost you reach; hiding it can cost you the channel.
In our work this is settled before anything is generated. In the brief we note which scenes are synthetic, who lends a face or voice and on what rights, so labels and releases are ready when the video goes live. If you want to set your channel up the same way, let’s talk.
Frequently asked questions
Can you monetise AI-made videos on YouTube?
Yes, as long as they are original and built on their own idea. YouTube excludes inauthentic content, meaning mass-produced or repetitive videos, from the Partner Program, not videos where AI is simply a production tool.
What happens if I don’t disclose an AI video on YouTube?
Since May 2026 YouTube can apply the label automatically when its systems detect undisclosed photorealistic AI. If the omission is consistent, the platform can remove content or suspend the channel from the Partner Program.
Does YouTube likeness detection recognise voices?
No, as of November 2026 the tool detects faces. When you file a removal request YouTube asks whether the video also copies your voice, but it cannot find matches based on voice alone.
Sources
- Improving AI labels for viewers and creators · YouTube Official Blog, 2026-05-27
- YouTube CEO Neal Mohan’s 2026 Letter: The Future of YouTube · YouTube Official Blog, 2026-01-21
- YouTube shifts generative AI labels to spots viewers will actually see · PPC Land, 2026-05-30
- YouTube’s AI deepfake detection tool is now available to all creators 18 and older · Engadget, 2026-05-16
- YouTube expands likeness detection to all users over 18 · Social Media Today, 2026-05-17
- YouTube Expands AI Likeness Detection to Celebrities · eWeek, 2026-04-22
- YouTube expands its likeness detection technology, which detects AI fakes, to a handful of top creators · TechCrunch, 2025-04-09
- YouTube prepares crackdown on 'mass-produced' and 'repetitive' videos, as concern over AI slop grows · TechCrunch, 2025-07-09
- YouTube clarifies monetization update on inauthentic, repeated content · Social Media Today, 2025-07-13
- YouTube is fighting AI slop with new monetization guidelines · BetaNews, 2025-07
Keep reading.
AI slop: what it is, why brands should care, how to avoid it
What AI slop is, why dictionaries and linguists made it the word of 2025, and how to tell a directed AI video from content churned out in bulk.
Generative engine optimization: how video gets your brand cited by AI
AI engines cite verifiable facts and brands mentioned on YouTube. What GEO research shows and what to do with your videos and transcripts.
AI Christmas ads in 2026: who uses AI, who refuses it and why
Prada draws by hand, Levi’s goes back to claymation, Dior faces AI accusations: Christmas 2026 shows what audiences punish and accept from AI.