An analytical review of the Flock Safety platform code has revealed a new development — the OS Investigate system, which fundamentally changes the approach to search operations. This AI tool allows law enforcement agencies to identify drivers based solely on their behavioral patterns and travel routes, completely eliminating the need to know a name or license plate number.
Architecture and Capabilities of OS Investigate
During a technical analysis of more than 450 files loaded upon authorization on the Flock website, it was possible to partially reconstruct the system's interface. A key feature is 69 ready-made AI query templates that an officer can adapt or create from scratch. The system integrates with surveillance camera data, police reports, 911 call logs, as well as commercial databases containing phones, addresses, and information about family relationships.
Of particular interest are 19 templates aimed at identifying behavioral anomalies, and 14 that operate without any identifiers. It is enough to set a geographic point and a time interval for the algorithm to begin searching for specified activity. This opens up opportunities for proactive surveillance, not just reactive investigation.
Algorithm for Identifying "Associations"
Another function of the system is the analysis of correlations between vehicles. The algorithm records license plates appearing at the same cameras within a two-minute interval of the target vehicle. With three or more matches and exceeding a confidence threshold, the vehicle is added to a list of potential "associations" — up to 20 results per query. This allows for building implicit social graphs without direct access to personal data.
Flock representatives emphasize that cameras do not identify drivers, and data is stored for only 30 days. Paris Lubel, an official company representative, noted that OS Investigate is a separate product for working with information already available to law enforcement, and it is in the development stage. However, the request justification form, as it turned out, requires only three characters — which calls into question claims of strict oversight.
Context and Growing Criticism
The Flock network covers more than 6,000 communities in the United States and processes about 20 billion license plate reads monthly. The Justice Institute has documented dozens of cases of abuse, including surveillance of romantic partners. In response, Flock points to logging of all queries and tools for detecting suspicious activity. Recall that in July, Congressman Thomas Massie proposed stripping federal funding from municipalities using such systems.
My analysis: This technology blurs the line between targeted investigation and total surveillance. Even if Flock declares data deletion, the very existence of an AI capable of building behavioral profiles without identifiers creates a dangerous precedent. The surveillance market is moving toward a model where privacy becomes not a right but a privilege — and the industry should think about strict regulatory frameworks before such systems become the standard.