Flock Testing New AI Tool: No License Plate Needed to Track People by Driving Habits

Police don't need a license plate, don't need a name, don't even need a case that's already occurred — just draw a boundary on a map, enter a time range, and type a sentence describing behavior, and the algorithm handles the rest. This is how Flock's new tool works, according to a report by tech outlet Wired, and it reveals the biggest gap yet between the company's public image and its actual capabilities.

Flock has long positioned itself as simply an automatic license plate recognition (ALPR) company, but outsiders have long suspected its cameras could do far more than just read plates. According to the Wired report, Flock internally developed new software under the codename Nightshift, now publicly known as OS Investigate. This system connects a camera network spanning 6,000 communities across the US, tracking vehicle movement within those communities. For example, if an incident occurs somewhere, the system can analyze vehicle activity near the scene to identify potential witnesses, then cross-reference those leads against police case files, 911 dispatch records, and commercial identity databases to further uncover a witness's name, address, family relationships, and even friendships.

The Wired report notes that this tool "can search for people matching a description within a map-defined area, based solely on a description of their appearance." In other words, officers don't even need a license plate, a name, or a specific case to begin with — they just need to provide a location, a time frame, and a behavioral pattern, and the algorithm does the rest on its own. The system features a built-in chatbot interface that lets officers type prompts directly, such as "Find the vehicle that appeared most frequently in [community] during [time period] over the past 14 days." Fill in the blanks and submit to run the search — and these prompts don't require the user to input any specific case at all.

Flock states that OS Investigate is currently only being tested with select law enforcement partners, and that its current functionality "may not reflect" the version that will eventually be sold. The company has not denied any of the capabilities described in the Wired report, only emphasizing that this is a separate product from its license plate recognition technology.

Between "Completely Insane" and "A Bit Uneasy"

Noel Pichardo, a former police officer invited by Wired to review the software and its prompts, said bluntly: "I don't know how else to describe it, but this sounds completely insane. I don't know who could argue against the fact that Flock is indeed tracking people." An anonymous active-duty officer said the feature for finding witnesses made him "a bit uneasy," but added, "As law enforcement, regardless of what the 'defund the police' crowd says, we have to use every tool available to us." Independent security researcher Buchodi took a more measured view, noting that "there's a lot of innocent, everyday behavior that could get flagged because of this, but it could also be quite useful for investigators."

ACLU attorney Chad Marlow zeroed in on the core risk: "An officer asking 'Did you see any pattern of crime' produces a completely different result than 'Find me a pattern of crime.' And they don't seem to have any restrictions on prompts at all — which essentially lets police guide the AI to do almost anything they want."

Flock has repeatedly faced accusations of misuse by officers in the past, including using cameras to surveil specific individuals and stalk ex-partners, prompting the company to add safeguards such as flagging abusive accounts and requiring searches to include a case number. As for what safeguards the still-in-testing OS Investigate might need, neither the Wired report nor Flock's current response offers an answer.

The unease surrounding Flock has already spilled into real-world action: some law enforcement agencies have ended their partnerships with the company, while some residents have taken matters into their own hands by vandalizing, removing, or stealing Flock cameras from the streets. In one small Minnesota town, every Flock camera in the area vanished overnight, and the local police department said they will not be reinstalling them.