A.I. News Flock is testing a new AI tool that tracks and identifies people based on their driving habits

[Brownie2019]

Content source

https://www.engadget.com/2240069/flock-testing-new-ai-police-tool-tracks-identifies-drivers/

This goes way, way further than keeping an eye on license plates.
Flock has long pretended to be a simple automated license plate reader (ALPR) company, despite overwhelming evidence that its cameras can track a lot more than that. Now there's a report that it has been developing new AI tools that can potentially locate people by how they drive, according to Wired.

The software was reportedly called Nightshift and is now going by the name OS Investigate. It draws from a network of cameras in 6,000 communities and logs the movements of drivers in those communities, Wired reports. This could be used, for instance, to find a witness to a crime by analyzing vehicle movements near where a crime occurred. Then, that information could be cross-referenced with police case files, 911 dispatch logs and commercial identity records to find the name, home address, relatives and even friends of the potential witness.

Wired says the tool can "search for people in an area drawn on a map based on nothing more than a physical description." In other words, an officer doesn't even need a plate, name or a crime to begin searching. They just have to provide a location, a time and a pattern of behavior. The algorithm does the rest.

Read More: Flock is testing a new AI tool that tracks and identifies people based on their driving habits - Engadget

 

[Bot]

Read More: Flock is testing a new AI tool that tracks and identifies people based on their driving habits - Engadget

Privacy and accuracy concerns

If the reporting is accurate, this would represent a significant expansion beyond conventional automated license plate recognition. An ALPR system generally records a vehicle, plate, time, and location; using movement patterns and contextual data to infer a person’s identity or locate witnesses is a much broader form of behavioral surveillance.

The main concerns include:

• False positives caused by incomplete camera coverage, shared vehicles, unusual routes, or inaccurate vehicle descriptions.
• The possibility of identifying innocent witnesses, bystanders, relatives, or associates who are not suspected of wrongdoing.
• Function creep, where data collected for locating vehicles is later used for person-based investigations.
• Data retention, access controls, auditing, and whether affected individuals can challenge incorrect records.
• Combining vehicle-location data with dispatch records, police files, and commercial databases, potentially creating highly detailed personal profiles.
• Insufficient transparency about the system’s training data, accuracy, safeguards, and law-enforcement approval process.

The phrase “based on nothing more than a physical description” is especially important. A description such as clothing, height, or apparent gender can be vague and can also introduce serious bias. A system producing a lead is not the same as establishing someone’s identity or involvement, so any result should require independent verification and should never be treated as conclusive evidence by itself.

I cannot independently confirm the specific capabilities or current name of the reported tool from the excerpt alone. The relevant questions are whether the tool is actually deployed, which agencies can use it, what data sources it accesses, how long records are retained, and what legal and technical safeguards apply. Those details should be established through the company’s documentation, public-records requests, agency policies, and independent testing.

Regardless of the branding, combining persistent vehicle tracking with identity databases deserves substantial public scrutiny and clear limits.