Clearview AI Is Testing an AI Tool to Automatically Dig Up People's Online Lives
Facial recognition company Clearview AI is developing an experimental tool called InquiryIQ that would use AI to automatically search the web for details about people identified by police. The company says it remains an unreleased prototype.
Clearview AI, the company known for scraping billions of photos from the internet to power facial recognition for police, is developing a new tool called InquiryIQ, according to reporting by Wired based on code found in files sent to visitors of Clearview's login page before sign-in.
InquiryIQ is designed to take over after an investigator runs a face-recognition search, automatically browsing web pages, analyzing images, and assembling a profile of a person's possible employers, aliases, associates, and physical traits. The interface suggests that entering a subject's age, gender, and race can help the AI make "smarter decisions" while searching.
One of the models tested to power the tool comes from SpaceXAI (formerly xAI), maker of the chatbot Grok, which has repeatedly drawn criticism for producing racist and extremist content. It remains unclear how the model would use demographic inputs, and Clearview declined to speculate.
Experts say such tools could speed up investigations dramatically, but also lower the barrier to broad, speculative surveillance of people police might otherwise never have looked into. Because generative AI can produce different results from the same starting point, it may also be hard to reconstruct why a system pursued a particular lead.
Clearview says InquiryIQ is only a prototype that has never been pitched or delivered to law enforcement customers and is not currently planned for release in its present form. CEO Amos Kyler says human review remains central: an officer must independently verify any finding before accepting it into a case file, and the tool is meant to surface leads, not determine facts.
Privacy researchers caution that this kind of human oversight can become a mere formality over time, as investigators grow accustomed to trusting automated output. In one past case involving Clearview data, an investigator accepted a match the system itself had flagged as "a less likely result," which later proved to be mistaken.
