Feed a stranger’s photo into Clearview AI InquiryIQ, and it can hand back their name, employer, phone number, and home address in minutes, no human analyst required. It’s a prototype reporters found buried in Clearview’s own website code this week, running on xAI’s Grok model. Clearview says it’s never been sold or used by police, but the find has privacy researchers worried about how close “find a face” and “build a dossier” have become.
Here’s the whole story in one skim:
- What it is: An AI agent that automates the identity research that normally follows a facial-recognition match.
- Powered by: xAI’s Grok model, layered on top of Clearview’s existing face-search database.
- Status: An internal prototype, first reported on September 10, 2026. Clearview says it’s never shipped to a customer.
- Who can use Clearview at all: Only vetted U.S. government and law-enforcement agencies, under a 2022 settlement with the ACLU.
- Why it matters: Federal procurement records show at least one agency already asked for this exact capability.
What Clearview AI’s InquiryIQ Actually Does
Regular Clearview facial recognition already does one job. It matches a photo against billions of images scraped from the open web, then shows where else that face turns up. InquiryIQ picks up right where that search ends.
Once a match comes back, the tool runs its own web searches on the person. It opens the pages it finds and applies a second round of facial recognition to any new photos on them. It keeps looping — cross-referencing names, workplaces, and addresses — until it has enough for what Clearview internally calls a “Candidate Graph.”
According to a report from Biometric Update, reporters spotted code referencing InquiryIQ inside files that Clearview’s own login page quietly sends to any visitor’s browser. The tool runs on xAI’s Grok — the same model family whose next release just got pushed back — rather than anything Clearview built itself.
The Candidate Graph: From One Face to a Full Dossier
The gap between a normal facial-recognition hit and an InquiryIQ result is the whole story here. One gives you leads. The other gives you a person. This table breaks down what changes when the research step gets automated.
| What You Get | Standard Face Match | InquiryIQ Candidate Graph |
|---|---|---|
| Output | A list of photos where the face appears | Name, address, phone, employer, aliases, social accounts |
| Human effort | An analyst researches each lead by hand | Fully automated, no manual research step |
| Speed | Minutes for the match, hours or days for follow-up | Minutes, start to finish |
| Data sources | Clearview’s own image database | Database plus live web search and secondary face matches |
That automation is exactly what makes people nervous. Removing the human researcher doesn’t just save time. It also removes the one point in the process where someone might pause and ask whether digging this deep is justified.
Why Experts Are Nervous About Grok Running the Show
It’s not just the automation that worries privacy researchers. It’s which AI is doing the automating. Grok has drawn repeated scrutiny for producing extremist and inflammatory outputs in unrelated contexts.
Legal experts quoted by Cryptonomist questioned whether a model with that track record should ever inform decisions tied to probable cause. A wrong facial-recognition match is already a documented problem — several people have been wrongfully arrested in the U.S. after bad matches. Layering in an AI agent that autonomously chases leads adds a second place where a strange model output could send an investigation toward the wrong person.
What Clearview Says — and What the CBP Paperwork Suggests
Clearview’s own position is straightforward. InquiryIQ “remains merely a prototype” that has “never been offered or delivered to any customer,” with no current plans to release it. The company says no law enforcement user has ever touched it.
That claim sits next to a less reassuring detail. Customs and Border Protection issued a Statement of Work in February 2026 seeking 15 Clearview licenses for Border Patrol intelligence staff. The stated purpose: “strategic counter-network analysis” — language that describes exactly the kind of profile-building InquiryIQ automates.
Clearview’s tools are already restricted to vetted government customers under its 2022 ACLU settlement. So this isn’t really about who’s allowed to buy access. It’s about how much one tool should be able to infer from a single photo.
Can You Do Anything to Protect Yourself?
There’s no setting that pulls your face out of Clearview’s database on demand. A few practical steps can still narrow your exposure.
- If you live in a state with a biometric privacy law, like Illinois or Texas, you may have a legal right to ask what data a company like Clearview holds on you.
- Set social profiles that show your face to private where you can. InquiryIQ’s whole method depends on finding more photos to chain together.
- Keep your browser and phone patched. Plenty of personal-data leaks start somewhere much simpler, like the Chrome zero-day AppAuxin covered last week.
None of this makes you invisible to a tool trained on billions of scraped images. It does shrink the amount of fresh material a system like InquiryIQ has to work with.
Frequently Asked Questions
What is Clearview AI’s InquiryIQ?
It’s a prototype AI agent that automates the research step after a Clearview facial-recognition match, using xAI’s Grok to search the web and compile a profile called a “Candidate Graph.”
Is InquiryIQ available to police or the public right now?
Clearview says it’s an internal prototype that has never been sold to or used by any customer, including law enforcement. There’s no confirmed release date.
What AI model powers InquiryIQ, and why does that matter?
It runs on xAI’s Grok. Grok has a history of producing extremist and inflammatory outputs in other contexts, which is why legal experts are uneasy about it feeding investigative decisions.
Can I find out if Clearview AI has my photo?
There’s no public self-service lookup. Residents of states with biometric privacy laws can submit a data request, though rights and response times vary by state.
Is Clearview AI legal to use in the US?
Under its 2022 ACLU settlement, Clearview can only sell access to vetted U.S. government and law-enforcement agencies — not to the general public or private companies.
The Bottom Line
A prototype that’s never shipped isn’t a reason to panic, but it is a preview. The gap between “match a face” and “build a stranger’s entire profile in minutes” is closing fast. A federal agency already asking for that exact capability tells you which direction this is headed. It’s worth watching closely, and worth treating Clearview’s “just a prototype” line as a snapshot of today, not a promise about next year.
Leave a Reply