YouTube broadens biometric likeness protection against AI deepfakes

YouTube is expanding its biometric deepfake detection system to all creators over the age of 18, broadening the platform’s efforts to combat AI-generated impersonation as synthetic media fraud spreads across politics, entertainment and professional sectors.

The video platform says that eligible creators will be able to sign up in the coming weeks, allowing them to identify videos containing synthetic or manipulated uses of their facial likeness. The goal is to keep their viewers from being misled, says the company.

The system uses selfie biometrics and government ID verification to detect synthetic or manipulated uses of a creator’s facial likeness in AI-generated videos, similar to YouTube’s Content ID system for copyrighted material.

Users enroll by submitting government-issued identification and a self-recorded verification video. The platform then alerts creators to matching content and allows them to request removal under YouTube’s privacy policies.

YouTube has been gradually expanding its deepfake detection tool to more users. In March, the feature became available to government officials, journalists and political candidates. The company said that it wanted to protect the people at the heart of important events, breaking news and civil debates.

The rollout later expanded to celebrities, allowing entertainers to combat unauthorized impersonation on the platform. YouTube is collaborating with talent agencies and management companies to provide the technology to entertainers.

The platform, however, continues to face biometric privacy questions. Media reports have noted that Google’s broader privacy policies leave open the possibility that creators’ biometric data could be used to help “train Google’s AI models and build products and features.”

The company says the biometric and identity data it collects for likeness protection is not used to train its algorithms.

Medical groups push biometric protections against AI “deepfake doctors”

After journalists, politicians and celebrities, the next group whose likeness could end up protected on the internet may be experts such as doctors.

Earlier this month, the American Medical Association (AMA) issued a policy framework for lawmakers and industry leaders aimed at preventing AI-generated deepfake “doctors” that impersonate physicians, manipulate the public and often endorse unproven treatments.

“When bad actors exploit a doctor’s identity, they undermine patient trust and can steer people toward harmful, unproven care,” says AMA CEO John Whyte. “We need strong action by federal and state lawmakers to protect physicians’ identities, ensure transparency and stop this fraud.”

The framework lists seven key steps for addressing AI-generated deepfake doctors. Among the recommendations is treating a physician’s name, image, voice and digital likeness as protected identity attributes requiring explicit opt-in consent before use in AI-generated content.

Consent must be specific and revocable, and content that falsely implies a physician’s endorsement or medical judgment should be classified as a deceptive practice. All such content must be clearly labeled and watermarked, with patients notified before interacting with an AI health professional.

The AMA also envisions shared liability across platforms, hospitals and AI vendors, with mandated takedown mechanisms for health-related deepfakes and a clear process for physicians to document misuse and seek remedies.

As AI-generated impersonation becomes more realistic and scalable, platforms, regulators and professional organizations are increasingly treating biometric likeness protection as a core safeguard against synthetic identity fraud and misinformation.

Article Topics

AI fraud  |  biometrics  |  deepfake detection  |  deepfakes  |  likeness detection  |  selfie biometrics