Clearview AI tested an agent that builds a dossier from a face match

Clearview AI, the facial recognition company built on billions of photos scraped from the internet, has been testing an AI agent that takes one face match and tries to reconstruct the person’s life around it.

What WIRED found

WIRED reported on September 10, 2026 that it found the prototype, called InquiryIQ, in files that Clearview’s login page sends to any visitor’s browser before they sign in.

After a law enforcement user runs a Clearview search, the tool is designed to search the web, browse pages, run face recognition on new photos it finds, and assemble a “Candidate Graph” of possible identities and associates. The graph lists addresses, phone numbers, employers, aliases, social media accounts and arrest histories. One of the models Clearview tested to power it came from xAI, the maker of Grok. WIRED reports that the interface also says supplying age, gender and race can help it make “smarter decisions.”

Clearview’s response

Clearview says InquiryIQ is an internal prototype that has never been pitched or shipped to customers and has never been used by law enforcement. The company also disputes that the tool amounts to an automated investigator.

Why it matters

Facial recognition was already contested because a single photo could put a name to a stranger. A system that keeps going, from name to address to employer to associates, changes the scale of what one search produces. It compresses hours of manual follow-up into minutes.

It also compresses mistakes. Face matches are probabilistic, and a wrong match at the first step becomes a wrong dossier by the last one, now populated with real addresses and real people. The prompt asking for age, gender and race adds a further question about how demographic inputs shape the results. Whether tools like this ever reach police will depend less on the technology than on the rules set around it.