The following guest commentary by Karl W. Bickel was published by The Baltimore Sun on August 25, 2026, and appears in Wednesday’s print edition. Karl spent his career in law enforcement and at the U.S. Department of Justice, and he writes here about why the rapid spread of Flock Safety’s license-plate-reader network deserves a harder look from the profession that uses it.
A cautionary perspective on Flock cameras
Law enforcement professionals who signed off on Flock Safety cameras did it for the reasons the company sells: to solve car break-ins, find abduction suspects and close cases that would otherwise go cold. Automated license-plate readers are effective, cheap compared with staffing and popular with communities looking for a crime-reduction measure. However, a growing number of community stakeholders are raising a different concern, not about whether the technology works but about what happens once it exists.
Flock’s cameras are more than a local tool. The company operates more than 120,000 license-plate-reader cameras under contracts with about 7,000 law enforcement agencies nationwide. A patrol officer running a plate in a small town in Ohio can, depending on sharing settings, pull results from cameras in California or Texas. A system built for local investigations became, almost by default, a national surveillance tool.
Flock has been consistent in its public messaging: ICE cannot directly access Flock cameras or customer data, and any federal access must be granted by a local customer in compliance with the law. That’s a real distinction, but it hasn’t reassured many community stakeholders. Reporting indicates that more than 4,000 lookups have been conducted at the request of the federal government for immigration purposes, effectively giving federal agents indirect access to a system they don’t have a formal contract for. In one documented case, an ATF analyst improperly used Richmond, Virginia’s Flock system to assist an immigration investigation after detainees escaped an ICE facility; Richmond police responded by cutting off federal access and confirming the employee involved no longer worked for the department.
That’s the pattern that worries detractors: not a formal ICE partnership but a side door that opens whenever one participating agency’s data-sharing settings are looser than another’s. A department that has never agreed to help enforce immigration law can still have its data searched, because the network doesn’t respect the boundaries individual agencies think they’ve set.
To its credit, Flock has responded to pressure. In January, it added a single-toggle option letting any agency disable federal data sharing, and the following month it introduced a post-login disclaimer flagging relevant state immigration and reproductive-care laws. Those are real safeguards. But they’re reactive, added after journalists and advocacy groups documented misuse, not built in from the start. That sequencing bothers people who’ve spent careers being told that data systems need privacy protections designed in, not bolted on after a scandal.
The backlash has been substantial. Cities including Mountain View and El Cerrito, California, have terminated their Flock contracts, and town councils in Rhode Island and elsewhere have rejected or removed the cameras, evidence that local officials, including some in law enforcement, no longer trust the assurance that data stays local.
For an official who has studied police history, the FBI’s COINTELPRO program from 1956 to 1971 is the cautionary tale that should be a warning. That program didn’t begin as a plan to harass civil rights leaders, anti-war organizers and journalists. It began as a counterintelligence effort against a genuine threat, the Cold War’s Communist Party. Surveillance infrastructure built for one purpose was gradually repurposed to monitor, disrupt and discredit lawful political activity, often with no judicial oversight and no meaningful accountability until a break-in at an FBI office in 1971 exposed the program.
The lesson isn’t that today’s police running license-plate-reading systems are secretly plotting political sabotage. It’s that infrastructure outlives intent. A camera network justified as a tool against carjackers doesn’t stay confined to carjackers; it becomes available for whatever the next administration, the next task force or the next data-sharing agreement decides to use it for. COINTELPRO shows what happens when the technical capability to track people’s movements exists before the legal and institutional guardrails do. Flock’s own audit trail has already surfaced officers using the system to track romantic partners and investigate abortions, which is precisely the kind of mission creep history predicts.
None of this means license-plate-reading technology has no place in policing. It means the profession has an obligation to demand what COINTELPRO-era reformers eventually forced onto federal intelligence work: hard limits on retention, real audit transparency, warrant requirements for cross-jurisdictional searches and enforceable consequences when an agency violates another’s sharing terms, not settings that can be silently overridden by a partner department three states away.
Those voicing these concerns aren’t anti-technology. They’re the ones who understand, better than most, how quickly a useful tool becomes an unaccountable one once the public stops watching how it’s used.
Karl W. Bickel (karlbickel@comcast.net), a Democratic candidate for Frederick County sheriff, is retired from the U.S. Department of Justice, was previously second in command of the Frederick County Sheriff’s Office and is a former assistant professor of criminal justice.
Read the commentary at The Baltimore Sun: A cautionary perspective on Flock cameras | GUEST COMMENTARY
