A tool called Raven can take a photo of a room and work out which property it was shot in, and police in two states already use it.
Raven, built by a company called Graylark, matches fixtures, fittings and decorative details in an interior photo against images gathered online. If a similar shot of the same room sits on the web next to an address, such as on a rental or property listing, Raven can return that address as a lead. It was previously called GeoSpy, which was first conceived to place photos by landmarks. The LAPD and the Miami-Dade Sheriff's Office in Florida have used it, mainly on human trafficking cases.
A demonstration video appears to show it pulling from Zillow, Vrbo and Airbnb. Airbnb is reviewing the situation. Zillow said scraping (bulk automated copying of web pages) its data violates its terms of service. Graylark's Daniel Heinen called the product "purpose-built AI for the people working to find answers and protect lives." His LinkedIn post highlights matching across "millions of short-term rental units worldwide" and does not mention privacy.
The closest precedent is Clearview AI, whose face-search tool was built on images scraped from public sites and sold to police. Raven follows the same basic pattern with rooms instead of faces. Its reach depends on listing sites, at least one of which says scraping breaks its rules, so those companies hold real leverage. Spreading beyond its first two customers would likely mean meeting regional privacy laws and striking deals with the platforms.
The gap that matters is between purpose and capability. Raven is pitched at trafficking investigations, but the same method works on any property whose address and interior photos are already online.