Flock's AI surveillance network is booming in the US, and not everyone wants it on their street anymore

By Zendoric · September 1, 2026.

The BBC has devoted a short piece to a phenomenon that has been brewing in the United States for some time: the rapid expansion of Flock Safety, the company that has built one of the country's most extensive surveillance networks through automatic license plate reader (ALPR) cameras enhanced with artificial intelligence. The available material for this particular piece is a very brief video note —a headline and little more—, so it is worth being cautious with the details and sticking to what is public knowledge about the case.

Broadly, Flock's model is simple, and that is why it has grown so fast: it installs cameras on streets and access roads, cross-checks every plate that passes against databases of stolen vehicles or wanted persons, and sells the service to local police forces and neighborhood associations as a low-cost security tool. That same design is what has set off the alarm among privacy advocacy organizations: a camera meant to locate a stolen car also builds, plate by plate, a record of anyone's movements, and that record can be shared or queried for purposes very different from those announced when it was installed. Hence why several U.S. city governments and communities have begun to review or question their contracts with the company, rather than simply renewing them.

This is not a problem unique to Flock or to license plate surveillance: it is the usual pattern by which AI applied to security is deployed before clear regulation exists on what data is stored, who can query it and for how long. It is the same lag we have already seen in offensive cybersecurity or in the algorithmic moderation of decisions: technical capability advances in months, governance rules take years. The outcome is predictable —fast adoption, then an isolated scandal, then a belated attempt to set limits— and it is exactly the kind of short-term friction that should not be downplayed.

Our reading is that the underlying debate is not whether these camera networks should exist, but who audits them and under what conditions. A tool that helps find a missing child or recover a stolen vehicle has real value; the same tool without access control, without traceability of who queries what and without limits on whom the data is shared with stops being public safety and becomes surveillance without oversight. The abundance and progress we defend as AI's horizon hold up only if they go hand in hand with governance based on evidence, not on security marketing or reactive panic. The Flock case, even if today's coverage is minimal, is one more gauge of a tension that will keep surfacing every time a company deploys AI sensors faster than institutions learn to oversee them.