Clearview Tests AI-Powered Police Investigations

See how generative AI could turn facial identification into broad online-personal-data searches.
30-Second TL;DR
What Changed
InquiryIQ is a previously unreported Clearview prototype
Why It Matters
The system could expand law-enforcement access to aggregated personal information and increase risks of profiling or mistaken identity. AI builders working with sensitive data should treat identity resolution and surveillance use cases as high-risk.
What To Do Next
Audit any identity-resolution workflow for consent, access controls, false-match rates, and documented law-enforcement authorization before deployment.
Key Points
- •InquiryIQ is a previously unreported Clearview prototype
- •It tested an xAI model to gather online information about identified people
- •Potential outputs include associates and social-media accounts
Deep Insight
Background and context from public sources — not the original article. 12 sources cited.
Enhanced Key Takeaways
- •InquiryIQ was uncovered through publicly accessible code embedded in Clearview AI's login page, rather than an official product announcement.
- •The system generates an autonomous 'Candidate Graph' that scrapes addresses, phone numbers, employment histories, aliases, arrest records, and known associates.
- •The prototype includes demographic input fields for age, gender, and race to guide the AI's web-gathering decisions.
- •Clearview stated that InquiryIQ remains an internal benchmark prototype that has never been sold, piloted, or deployed to law enforcement clients.
- •Clearview AI's underlying scraped facial recognition database has expanded from 3 billion images in 2020 to over 70 billion images.
Competitor Analysis
- Primary Ingestion Modality
- Facial recognition & web scraping (70B+ images)
- Investigative Output
- Automated Candidate Graph (associates, social media, dossiers)
- Target Market
- Law enforcement & federal agencies
- Primary Ingestion Modality
- ALPR (License plate recognition) & public records
- Investigative Output
- Pattern-of-life dossiers and suspect location mapping
- Target Market
- Municipal police departments & local law enforcement
| Company / Product | Primary Ingestion Modality | Investigative Output | Target Market |
|---|---|---|---|
| Clearview AI (InquiryIQ) | Facial recognition & web scraping (70B+ images) | Automated Candidate Graph (associates, social media, dossiers) | Law enforcement & federal agencies |
| Flock Safety (OS Investigate) | ALPR (License plate recognition) & public records | Pattern-of-life dossiers and suspect location mapping | Municipal police departments & local law enforcement |
Technical Deep Dive
- Model Integration: Integrated with xAI's Grok large language model architecture to parse unstructured web data, conduct secondary queries, and synthesize personal dossiers.
- Candidate Graph Pipeline: Executes recursive web and image searches, navigates external webpages, and re-applies facial recognition to construct a dynamic, relational entity graph.
- Demographic Heuristics: Ingests user-specified demographic parameters (age, gender, race) to narrow automated search spaces and refine entity resolution heuristics.
- Verification Checkpoint: Implements a client-side attestation prompt requiring users to certify independent factual verification prior to committing records to permanent files.
Future ImplicationsAI analysis grounded in cited sources
Timeline
- 2020-01Clearview AI comes to widespread public attention with a 3-billion-image database
- 2022-05Clearview settles ACLU lawsuit under BIPA, restricting commercial sales in the US to government and police
- 2024-03Clearview's indexed database scales to tens of billions of scraped web images
- 2026-09WIRED reveals unreleased InquiryIQ prototype testing xAI models to automate personal dossiers
Sources (12)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
Weekly AI Recap
Read this week's curated digest of top AI events →
AI-curated news aggregator. All content rights belong to original publishers.
Original source: Wired AI ↗
This is a summary, not the original. Read the source, or get the weekly briefing.
The weekly digest
One email a week. Unsubscribe anytime.