How YouTube content creators can stop others from using their likeness to create deepfakes

The threat of AI-generated deepfakes and synthetic content that impersonates a creator’s face or voice has become one of the most serious issues facing the digital world. In response, YouTube has officially launched its Likeness Detection technology, expanding it to eligible members of the YouTube Partner Program (YPP).

This new AI-driven tool empowers creators to proactively find and submit requests for the removal of unauthorised videos that are stealing their digital identity, restoring a critical layer of control in an era of rapidly advancing generative AI.

This technology acts as an automated digital guardian, working similarly to YouTube’s established Content ID system, but focused entirely on protecting a person’s identity instead of just copyrighted video or music.

For creators, this means a systematic way to combat the use of their image for fake endorsements, scams, or the spread of misinformation, which can severely damage their brand and relationship with their audience.

The system works by using a secure, verified reference of your face to scan the millions of videos uploaded to the platform daily. When a potential match is found, which could be a simple re-upload of your content or an AI deepfake of your face or voice, it is highlighted on a private dashboard.

This moves creators from a reactive position to a proactive one, giving them the power to immediately review the content and request a takedown under YouTube’s privacy guidelines.

It’s important to note that the setup process requires submitting sensitive information like a face scan and a government ID to ensure that only the rightful creator can use the tool to file claims.

Step-by-Step guide to activating YouTube likeness detection

If you are an eligible member of the YouTube Partner Program (YPP), the Channel Owner, or a Manager over 18, you can set up the Likeness Detection tool in a few simple steps via YouTube Studio.

Step 1: Access the detection dashboard

Open YouTube Studio on a desktop browser.
In the left-hand menu, navigate to the content detection tab.
Select the Likeness sub-tab and click “Start now” to begin the onboarding process.

Step 2: Agree to the terms

Review the terms of service for the Likeness Detection feature.
You must agree to YouTube using biometric technology (your face and voice data) to scan the platform for unauthorised matches of your likeness.

Step 3: Complete identity verification

This step is mandatory to create a trusted reference template and prevent fraudulent claims.
Follow the on-screen prompts to complete the Google identity verification, which requires uploading a clear image of your official government-issued ID and recording a brief video selfie of your face as instructed.

Note: Verification may take up to five days to be approved by YouTube.

Step 4: Review detected videos

Once verified, the AI system begins scanning new uploads.
Periodically check the Content detection > Likeness > For review tab in YouTube Studio.
This dashboard will list any videos where your likeness has been detected.

Step 5: Take action on a match

Click “Review” next to a flagged video to examine the content, channel, and context.
You must then decide on the appropriate action:
- Request content removal: Select this if you believe the video uses your likeness in a misleading or harmful way (e.g., an AI deepfake making a fake endorsement) and violates YouTube’s privacy guidelines. You will complete a privacy complaint form.
- Take no action/Archive: Select this if you are okay with the content, if it is not actually you, or if it is legitimate footage (like a brief news clip or fair-use parody). The video will be moved to the Archived tab.

While the system is powerful, it may not catch every instance, and there may be occasional false positives. However, this Likeness Detection tool marks a critical and necessary advancement in providing creators with the tools to actively protect their professional identity against the rising wave of sophisticated AI impersonation.

Source: Technext24