AI Police: Flock Relies on Movement Patterns for Surveillance

Imagine the police could identify suspects solely by how they drive—no license plate, no name, just their driving style. Sounds like science fiction? For many U.S. authorities, it’s becoming reality. The company Flock Safety has developed a system called "OS Investigate" that does exactly this: it tracks vehicles based solely on their movement patterns and cross-references them with suspect profiles.
I have to admit, the concept is impressive—and somewhat unsettling. Flock Safety, a provider of surveillance cameras already deployed in over 4,000 U.S. cities, claims its new AI software doesn’t just scan license plates but also analyzes subtle driving maneuvers—how someone accelerates, takes turns, or even parks. This data is stored in a vast database and matched against suspicious profiles. In theory, police could identify criminals before they even have a name or face.
The technology is constantly evolving—much like Netflix, except instead of your film preferences, it analyzes your driving behavior. Flock Safety is confident, citing promising early tests. But here’s the catch: the algorithms are secret. No wonder privacy advocates like the EFF are sounding the alarm. "If police can monitor vehicles based solely on movement patterns without concrete suspicion, this quickly becomes mass surveillance," warns the civil rights organization—and they’re not wrong. The risk that "efficient law enforcement" turns into "automated round-the-clock monitoring" is real.
Legally, the situation in the U.S. is a patc



hwork. Some states already use "OS Investigate," while others debate whether it complies with the Fourth Amendment—the ban on unreasonable searches. In Portland, Oregon, a lawsuit against Flock Safety is underway, with citizens fearing their right to privacy is being violated. Europe would be far stricter: under the GDPR, such a system would likely be illegal. But let’s be honest—how long will it be before this technology arrives there too?
The debate isn’t new. Police agencies worldwide already use facial recognition, predictive policing, and other AI systems to anticipate crime. The question remains the same: where does legitimate investigation end, and mass surveillance begin? Flock Safety insists its system is only used within the scope of specific investigations. Yet history shows that technologies tend to expand. What’s used today for serious crimes could become standard procedure for minor traffic offenses tomorrow.
I wonder: how much surveillance are we willing to accept in exchange for feeling safer? Sure, if it helps solve burglaries or car thefts more quickly, it sounds good at first. But where do we draw the line? And who ensures that this line isn’t crossed?
One thing is certain: the development won’t stop. "OS Investigate" is just the beginning. Society must decide how much it wants—and, more importantly, how much it’s willing to give up. Otherwise, the idea of smart policing could quickly turn into a surveillance tool that tracks us all. And that would be a bit too much of a good thing.