Scan of 1 million web assets finds thousands of suspected AI likeness uses

New pilot data from Los Angeles-based indieme.ai found thousands of suspected unauthorized or synthetic likeness uses across more than one million public web assets after verifying about 15,000 talent records. The results underscore the gap between declaring AI consent and enforcing it as lawmakers and rights holders push for stronger protections.

Why it matters: - The pilot data shows how quickly public web content can outpace consent systems for talent likenesses. - The findings highlight a practical problem for creators, agencies and rights holders: stating AI-use preferences does not automatically block misuse, prove authenticity or trigger takedowns. - The results arrive as governments and industry groups move toward clearer rules for digital likeness rights and nonconsensual AI content.

What happened: - indieme.ai scanned more than one million publicly available images and assets across the open web. - The scan flagged thousands of instances of suspected unauthorized or AI-generated likeness use. - The data came from completed pilot programs with State Management and NLE Group International. - Those pilots verified about 15,000 talent likeness records. - The pilots followed the June 24, 2026 launch of the Human Consent Registry at the European Parliament by actor Cate Blanchett. - RSL Media developed the free registry tool.

The details: - The Human Consent Registry lets individuals declare how AI systems may use their name, face, voice and likeness. - The registry includes options to permit use, permit use with conditions, or prohibit use. - Blanchett said, "Your identity is your IP in the age of AI," according to Euronews. - indieme.ai said talent likeness data was encrypted at the point of capture to create a verified reference record. - The company's AI Media Detector then scanned public web content against those records and flagged suspected unauthorized or synthetic uses for review. - Authorized images are tagged with C2PA, a content provenance standard supported by Adobe, Microsoft and Intel. - indieme.ai said its Likeness Library is designed to let verified talent likenesses be licensed for commercial use, with fees and a share of revenue going to talent. - The company said no licensing transactions have occurred through the pilots. - Dion Johnson, founder and CEO of indieme.ai, said the registry sets terms while the pilot data shows the volume of material already circulating that needs to be checked against those terms. - Johnson said thousands of cases required review across roughly 15,000 talent records.

Between the lines: - The pilot suggests a consent declaration alone is not enough to police AI likeness misuse at scale. - The workflow described by indieme.ai pairs verified capture, detection and provenance tagging, which points to a broader shift from policy statements to enforceable infrastructure. - The legal backdrop is tightening, with new rules and proposed federal rights pushing platforms and talent businesses toward more formal controls. - Early case examples cited in the release reflect how likeness abuse is already showing up in explicit content and scam advertising.

What's next: - indieme.ai said it plans to expand platform availability to additional talent representation firms after the pilot programs. - Wider adoption would likely test whether the system can scale beyond early partner firms and support licensing, review and enforcement workflows. - The broader market is still waiting for clearer national standards on consent, provenance and compensation for AI use of likeness.

The bottom line: - The pilot data suggests the AI likeness problem is already large, and the next battle is not just consent but enforcement.