Flock Safety Cracks Down on ALPR Abuse With Mandatory Audits and Shorter Data Retention
Flock Safety is rolling out mandatory oversight controls and shortening its default license plate data retention window from 30 days to seven, responding to a pattern of police officers across the United States using its automated license plate reader (ALPR) systems to stalk ex-romantic partners and others.
ALPRs are networks of cameras that capture and store license plate numbers along with the time and location of each sighting. Police use them to track vehicles tied to investigations, but the systems also log data on millions of ordinary drivers who are not suspects.
CEO Garrett Langley, in an interview with The Verge, delivered a mea culpa and said he changed his mind on the responsibility Flock bears for how law enforcement uses its tools. Langley said Flock got the oversight question wrong and decided to make its previously optional Audit Assistance feature mandatory. He also said more officers will be fired and more officers arrested for abusing police power.
The changes come after a string of documented abuse cases. A police chief in Braselton, Georgia, searched his ex-girlfriend's license plate more than 500 times over seven months, as reported by The Washington Post. California prosecutors alleged that a former Riverside County Sheriff's Office deputy used Flock data to stalk and harass his ex-girlfriend, looking up license plate information on men she was dating. The Institute for Justice, a nonprofit, has tracked reports of Flock and other ALPR abuse dating back to 2021. Officers have been losing their jobs over misuse of the technology, and advocates argue it is too easy to abuse, as Mother Jones reported in July.
Audit Assistance, which launched as a voluntary feature in April 2026, flags abnormal search behavior and immediately locks out suspicious users until an administrator reviews their activity. Think of it as a fraud-detection system for database queries: if someone runs hundreds of lookups on a single plate, the system freezes their access and routes the activity to a supervisor. Flock will now require it for all customers. The company will also require all users to enter case codes to run searches. Administrators can override the case-code requirement for emergencies such as tracking a missing child, and those overrides will be automatically flagged for review.
On data retention, Flock is moving its default ALPR data window from 30 days to seven, bringing it closer to limits long suggested by critics including the ACLU. Previously, Flock marketed its automatic 30-day deletion as a compliance tool to reduce risk, enforced by annual SOC 2 Type II audits (a standard third-party security assessment) and hard-deletion on a rolling schedule. Police departments can still choose to lengthen the retention period or use a new "Evidence Mode" that preserves searches tied to a specific case number.
Flock is also introducing granular data-sharing controls. Localities using the platform can now set per-agency sharing limits, such as allowing another agency to search their data for a homicide suspect but not for immigration enforcement. Previously, sharing operated on an all-or-nothing basis.
Langley had previously argued that lawmakers bore responsibility for crafting policies requiring audit systems or other limits on ALPR use. Little such regulation has materialized. That stance has now shifted: Flock is imposing the controls itself rather than waiting for legislative mandates.
The policy reversal carries weight because it concedes a structural problem that privacy advocates have identified for years. ALPR networks aggregate location data on millions of drivers who are not suspects, and the access controls around that data were, until now, largely left to the discretion of individual police departments. Mandatory audit-lockout and case-coded searches introduce friction at the point of query, which is where the abuse actually occurs. Whether these controls are sufficient depends on implementation rigor, administrator vigilance, and whether departments that want to circumvent them can simply configure looser policies at the local level.
Langley also appeared on The Drive's podcast in late July to discuss a separate controversy: errors in Flock's AI-powered camera network that led to a wrongful police stop of the podcast's host. That incident broadened the public debate beyond stalking abuse to include accuracy and false-positive concerns in Flock's vehicle identification systems.
Taken together, the changes address two distinct failure modes in ALPR deployment: human misuse of access privileges and the volume of data retained on innocent drivers. The seven-day default retention is a meaningful reduction. It narrows the window during which data on non-suspects exists in a searchable form and limits the blast radius of any single abuse incident. The case-code requirement and automatic override flagging create an audit trail where previously there may have been none.
What remains unresolved is enforcement. Flock can mandate features and set defaults, but police departments configure their own systems. Departments that opt into longer retention periods or override case-code requirements for broad reasons can still create conditions ripe for abuse. The granular sharing controls help, but they too depend on local administrators making sound policy choices. Flock has moved the floor, but the ceiling is still set locally.
Langley's prediction that more officers will face consequences is consistent with the pattern documented to date. The abuse cases that have surfaced did so because they were discovered, often after prolonged periods. More rigorous auditing will likely surface more of them. That is the mechanism working as intended, even if the resulting headlines are uncomfortable for the company and its customers.
