Flock rolls out new plate-reader safeguards

Flock tightens license-plate camera controls, new seven-day default retention meets opaque abuse-detection tool, auditing marketed as safeguard while methodology stays undisclosed

Flock, the surveillance technology company behind automated license plate reader camera networks, announced new policies and tools on Thursday that it says will curb misuse by law enforcement customers, according to TechCrunch. The company is now recommending that police departments retain data for seven days instead of 30, while offering an “Evidence Mode” that allows longer retention in exceptional cases if a specific case number is provided.

Flock also says customers can set limits on data-sharing with other departments based on the type of offense being investigated. These configuration changes are aimed at a recurring problem for plate-reader systems: once a network exists, the easiest “feature” is simply more access—more time, more agencies, more searches—because the cost of over-collection is pushed onto the public while the operational benefit accrues to investigators.

The headline product in Flock’s announcement is a tool it calls “Audit Assistance,” which the company says detects abnormal activity and flags it for administrator review. Flock launched Audit Assistance in April and says more than one-third of customers have enabled it; the company plans to require it for all customers by the end of the year. The tool can also lock out flagged users until an administrator intervenes.

What Flock has not provided is the sort of detail that would let outsiders judge whether the system is a meaningful control or a compliance checkbox. TechCrunch reports that the company has not explained how Audit Assistance works, what it is trained on, or how often it produces false positives and false negatives. When asked whether the tool uses machine learning or AI, Flock told TechCrunch it does not; it described Audit Assistance instead as a data tool that flags atypical search patterns for further review. One example that can trigger a flag is a user searching the same plate using multiple different case codes.

Flock executives and customers describe the system in broad terms. Ashley Haber, the company’s head of trust and compliance, said it can flag odd search history from a specific user. Co-founder Paige Todd said it can surface unusual patterns early, such as a user repeatedly searching the same plate over an extended period. Major Patrick Krieg of the Dunwoody, Georgia Police Department described it as an algorithm that notifies them of “any type of bias,” a claim that implies a much stronger capability than the company’s own description.

Critics argue that without transparency the company’s accountability pitch is impossible to evaluate. Chad Marlow, senior policy counsel at the American Civil Liberties Union, told TechCrunch there is no evidence the tool works consistently to address abuse. Marlow said Flock has pointed to arrests, firings and discipline of officers who misused the system as proof that auditing works, but that without knowing the overall scale of misuse those anecdotes do not establish effectiveness.

Flock is asking agencies to collect less data by default while selling them an internal alarm that decides when a user’s searches look strange. The company is also asking the public to accept that the alarm is reliable without showing how it rings.