In 2020, **Clearview AI** drew international scrutiny for amassing over 3 billion photos from the internet to power its facial recognition services for police and security professionals. Now, the company is experimenting with a new AI-driven approach, potentially enabling law enforcement to build extensive profiles about individuals, detailing their associates, residences, and online footprints. WIRED has uncovered details about **InquiryIQ**, an unreleased, experimental AI 'analyst assistant' developed by Clearview. This tool is designed to take initial details from a Clearview facial recognition search and autonomously scour the web, analyzing webpages and images to construct a profile that could include possible employers, aliases, associates, and physical characteristics of a person under investigation. Code analyzed by WIRED reveals that **InquiryIQ**'s interface suggests that providing age, gender, and race can help the system make 'smarter decisions' during its searches. Intriguingly, one of the models tested to power these decisions originates from **SpaceXAI**, the **Elon Musk** company behind **Grok**. **Grok** has faced criticism for racist, extremist, and inflammatory outputs, raising questions about how demographic inputs might influence **InquiryIQ**'s results. Clearview has not offered specific details on this integration. Experts warn that tools like **InquiryIQ** could fundamentally reshape police investigations, drastically reducing the time required for detective work. While this could accelerate crime solving, it also lowers the barrier for 'fishing expeditions,' enabling scrutiny of individuals who might otherwise never have been investigated. Furthermore, the generative nature of AI could make it difficult to trace the system's reasoning for pursuing specific leads. **Amos Kyler**, **Clearview AI**'s CEO, states that **InquiryIQ** is merely a prototype, has never been pitched or shipped to customers, and is not currently slated for release in its present form. Kyler disputes the characterization of the tool as an automated investigator, describing it as a limited means to automate web searches detectives already perform. He clarified that the **SpaceXAI** and other model options were for internal testing and comparison, not for police selection. "No law enforcement user has ever used it, period," Kyler affirmed. WIRED's discovery of **InquiryIQ** mirrored its earlier uncovering of **OS Investigate**, an AI-powered search tool under development by **Flock Safety**, by analyzing files sent to any visitor's browser from **Clearview AI**'s login page. These files contain extensive code and user interface text—instructions, warnings, and feature descriptions—that detail **InquiryIQ**'s design and intended functionality. They explicitly state that **InquiryIQ** is built to "automatically discover and enrich personal data from web sources." **Clearview AI** maintains that the prototype's apparent completeness should not be mistaken for a product nearing release, attributing it to the accelerated development cycles afforded by modern AI tools. Founded in 2017 with an early investment from **Peter Thiel**, **Clearview AI** initially operated with little public oversight, quickly signing up police departments nationwide. A 2020 **HuffPost** investigation highlighted founder **Hoan Ton-That**’s ties to the far-right. The **New York Times**' 2020 exposé on **Clearview AI**'s scraping of over 3 billion images from platforms like **Facebook**, **YouTube**, and **Venmo** catapulted the company into the spotlight as one of the most controversial surveillance firms. Despite demands from tech companies, lawsuits, and regulatory investigations, **Clearview AI** continued to expand its database, which now reportedly exceeds 70 billion images and serves over 2,000 law enforcement agencies. **Ton-That** stepped down as CEO in December 2024. **Kyler**, who assumed the CEO role last October, emphasizes a more deliberate and technically focused approach, stressing controls, auditing, and oversight. He frames the company's current focus as "refinement" and "ensuring that the product hits the kind of expectation of integrity that our customers expect." ### Signals and Noise For years, **Clearview AI** maintained that its role was limited to providing facial recognition leads, leaving further investigation to law enforcement. **InquiryIQ** appears designed to integrate some of this investigative work directly. As described by **Clearview AI**, **InquiryIQ** would activate after an investigator has conducted a facial recognition search and identified relevant details. These details, along with information such as age, gender, race, hair color, and eye color, can be added to a profile, which the interface indicates helps the AI make 'smarter decisions' during its searches. Kyler explained that **InquiryIQ** was conceived to test the validity of information identified by an investigator. The system would take a 'factoid,' conduct searches, and determine its relevance. The reviewed code describes **InquiryIQ**'s ability to perform web and image searches, browse webpages, and apply facial recognition to discovered photographs. As it processes information, the system is designed to construct a 'Candidate Graph' of potential identities and associates, while populating the subject’s profile with possible addresses, phone numbers, employers, social media accounts, arrest history, and aliases. **Clearview AI** is not alone in automating public information trawling. Other intelligence platforms, such as **ShadowDragon’s SocialNet**, **Penlink’s Tangles**, and **Fivecast**, already assist investigators in uncovering aliases, associates, and mapping digital footprints. **Andrew Guthrie Ferguson**, a George Washington University law professor specializing in AI and policing, likens this automated investigation to "digital rummaging." He notes that these tools can consolidate disparate online clues into a comprehensive profile. **Woodrow Hartzog**, a Boston University privacy scholar, highlights that the labor involved in traditional investigative work historically acted as a practical check on surveillance. By eliminating this friction, tools like **InquiryIQ** could dismantle existing privacy safeguards. "The privacy protections we have in place right now were mainly built in a world that assumed a certain amount of friction in the ability of governments to collect information about people," Hartzog stated. "There are a lot of rules we never had just because we never needed them—because there were these practical barriers to following everyone around." Upon completion of an **InquiryIQ** search, the interface is designed to present the officer with the identities, connections, and other details uncovered by the system, which the officer can then accept or reject.