Flock Safety is an Atlanta-based company that makes AI-powered automated license plate readers (ALPRs). It operates a network of approximately 120,000 cameras across the United States, contracted with local law enforcement agencies to help solve crimes, locate stolen vehicles, and find missing persons. By those operational metrics, it has been remarkably effective. By its own account, it helped locate 10,000 missing people last year and assisted in more than one million criminal investigations.
It is also now facing a Senate investigation, a class-action lawsuit, a growing movement of communities demanding removal of its cameras, and the kind of press coverage that reframes a company's entire brand narrative. More than 50 cities and counties cancelled their Flock contracts in 2026 alone — the highest cancellation rate in the company's history. July 2026 set the single-month record for contract terminations.
The operational product still works. The enterprise risk management did not.
What Actually Happened
The immediate trigger was federal data access — specifically, evidence that ICE, CBP, and other federal immigration enforcement agencies were querying Flock camera data through local law enforcement portals, in some cases explicitly prohibited by state sanctuary laws and local policy.
Two California communities facing the same public pressure produced opposite outcomes. Grass Valley — after a 60-day review and community input — suspended all cameras and cancelled its contract, with Police Chief Gammelgard saying "things have changed since that time, rapidly." Woodland — facing the same community concerns — held a city council presentation defending the technology and maintaining its contract. Both decisions were made by elected governments responding to the same underlying risk. The difference is in how each community weighed reputational risk, legal exposure, and stakeholder trust against operational benefit. That weighing process is enterprise risk management — and in both cases, the community forced it because the company had not.
The ERM Failure: What Flock Should Have Seen Coming
Flock Safety did not create these risks accidentally. It built a product that was genuinely useful, deployed it rapidly across thousands of jurisdictions, and grew to a scale — 120,000 cameras, 120,000 data portals — that made governance extremely difficult. That scale also made the company's exposure enormous. Every new camera was a new stakeholder relationship, a new data-sharing agreement, and a new potential point of failure.
The core ERM question is not whether Flock knew about data-sharing risks. It is whether those risks were formally identified, assessed, and treated before the company scaled — and whether the treatment was adequate to the scale achieved.
Here is what a proper ERM risk identification process should have surfaced — and what Flock appears to have underweighted at every stage of its growth:
The Federal Data-Sharing Dimension
The issue that turned a manageable privacy debate into a national crisis was the federal data-sharing question — specifically whether Flock's platform enabled ICE and other federal immigration enforcement agencies to access data collected by local police in jurisdictions that had explicitly prohibited such access.
Flock's own timeline of remediation tells the story of a company reacting to risks it had not fully anticipated. Federal agencies were removed from its national search network in August 2025. A toggle allowing agencies to disable all federal sharing was introduced in January 2026. A post-login disclaimer aligning with federal CJIS requirements was added in February 2026. Each of these controls is reasonable. Each of them came after the exposure, not before.
Reactive controls are better than no controls. But enterprise risk management is designed to identify exposures before they become crises — not to document the remediation steps taken after. The sequence matters. A company that builds a 120,000-camera surveillance network deployed in thousands of jurisdictions under thousands of different legal frameworks, with federal access enabled by default, and then introduces the first meaningful federal access controls years into operation, has inverted the risk management process. The exposure was built first. The governance followed.
What Proper ERM Would Have Looked Like
This is not hindsight criticism. The risks Flock is now managing were identifiable — and in some cases actively identified by outside observers — years before the current crisis. The ARM™ 400 risk identification framework asks organizations to scan their environment for emerging risks, map their stakeholder universe, and identify exposures that exist at the intersection of their business model and external forces beyond their direct control. Applied to Flock's situation circa 2022 or 2023, it would have surfaced every one of the six risk categories above.
A proper stakeholder risk assessment would have identified organized community opposition as a material business risk — not just a communications challenge. The DeFlock movement was already mapping cameras publicly. The ACLU was already investigating. Privacy advocates were already attending city council meetings. The signal was there.
A legal risk architecture appropriate to a multi-state deployment would have built state-specific compliance controls into the platform design — not added them in response to lawsuits. California's sanctuary laws, Oregon's sanctuary statutes, and state-specific privacy regulations were known quantities. A product deployed in all three states needed state-aware access controls from the start.
A governance framework for customer conduct would have recognized that Flock's reputational risk is inseparable from its customers' conduct. Every officer arrested for ALPR misuse becomes a Flock story. Contractual governance, mandatory audit requirements, and real consequences for misuse are risk treatments — not just good customer service.
An emerging risk horizon scan would have identified the federal data-sharing question as politically explosive years before it became a Senate investigation. The intersection of AI-powered mass surveillance and immigration enforcement, in a politically polarized environment, was not a subtle risk. It was a foreseeable collision between a product design decision and a national political fault line.
The Grass Valley Lesson for Every Organization
Grass Valley Police Chief Gammelgard said something worth examining carefully when he announced the suspension: "More than one thing can be true at the same time, while time and information can change the calculus."
That is a precise description of how emerging risk works. The original decision — adopting Flock cameras in 2021 — was reasonable based on the information available at the time. The cameras worked. They solved crimes. The community benefit was real. What changed was not the technology. What changed was the risk environment: the evidence of federal data access, the pattern of police misuse, the legal ambiguity, and the community's assessment of the tradeoff between safety benefit and civil liberties exposure.
The ARM™ 400 curriculum describes this dynamic explicitly: the risk management process is not a one-time assessment. It is a continuous cycle of scanning, identifying, analyzing, treating, and monitoring — with special attention to how the environment around an existing decision changes over time. Grass Valley did that re-assessment when the community forced it. A mature ERM framework does it before the community has to.
Every organization deploying AI-powered technology that collects data on individuals — whether cameras, sensors, employee monitoring systems, or customer analytics platforms — faces a version of the risk Flock is navigating. The questions are the same: Who else can access this data? Under what legal frameworks? With what consent? With what governance controls? And what happens when the community that hosts your technology decides the tradeoff is no longer acceptable? Those questions are not technology questions. They are enterprise risk management questions. They should be answered before deployment — not after the Senate investigation opens.
As of this writing, the Senate investigation is active, the class-action lawsuit is proceeding, and Flock is conducting community meetings across the country in what amounts to a reactive stakeholder engagement campaign — the kind that an ERM-informed proactive engagement strategy might have prevented from being necessary at all. The company has real tools, real benefits, and a real reputational crisis. All three are the result of the same underlying decision: to scale faster than the governance framework that should have accompanied the scale.
Reputational Risk, Legal Risk, and Emerging Technology Risk Are Core ARM™ Topics
The Flock Safety case illustrates every risk category covered in the redesigned ARM™ 402 curriculum — and the risk identification framework from ARM™ 400 that should have prevented it.
Sources: Axios Atlanta, "Flock's license plate readers fuel privacy concerns" (September 1, 2026); CBS Sacramento, "Grass Valley police indefinitely suspend Flock license plate cameras" (August 2026); The Union, "Powerful tools require powerful oversight: Grass Valley under pressure to cancel Flock surveillance contracts" (May 29, 2026); Union-Bulletin, "California's Woodland police defend Flock license plate cameras amid privacy concerns" (September 2, 2026); FOX40, "Grass Valley PD cancels Flock Safety contract over legal uncertainty" (August 2026); The Bend Source, "Federal immigration officials made 279 queries into Bend's Flock Safety data in its first three weeks" (May 2026); Immigration Policy Tracking Project, "Reported: ICE using automated license-plate-reader cameras for immigration enforcement" (August 2026); Senator Josh Hawley letter to Flock Safety CEO (August 26, 2026); Flock Safety blog, "Does Flock Share Data With ICE?" (January 2026); Gibbs Mura v. Flock Safety, San Francisco Superior Court (February 2026). This post reflects publicly available information as of September 3, 2026.
