While the company previously claimed its technology could not track individuals, this system features 69 prewritten prompts designed to help officers find witnesses, map the homes of frequently arrested residents, and identify a driver’s "associates" based on vehicles that travel together.
The development of this AI infrastructure marks a significant shift in law enforcement capabilities, moving from reactive searches for known stolen cars to proactive "fishing expeditions" based on behavioral patterns.
Privacy experts and legal analysts express concern that the system allows police to bypass traditional warrants by casting suspicion on drivers based on how often they visit specific locations, such as retail stores or banks.
This creates a competitive and civil liberties tension where automated "agentic policing"—using AI agents to execute multi-step investigative tasks—could establish a new standard of suspicion derived from algorithms rather than legal precedents.
Flock is currently testing the software with a small group of law enforcement partners and states that its final capabilities may change before a wider release.
The system’s technical mechanism relies on a chat-based interface where officers can enter natural language prompts to generate dossiers on individuals without needing a specific name or plate number to start.
As the company faces mounting bipartisan political pressure and reports of officer misuse, it plans to implement new safeguards by the end of the year, including automated monitoring for abnormal searches and mandatory case codes for all queries.