Former Pentagon AI Director Warns Flock Camera Networks Risk Unchecked Surveillance

A former senior Defense Department official's public warning about AI-enabled license plate readers raises concrete questions about federal oversight standards that do not yet exist in statute.

Mark Beall, who served as Director of AI Policy at the Department of Defense, stated publicly on September 5, 2026, that expanding deployment of Flock Safety camera networks across American communities could produce conditions resembling state-level mass surveillance without enforceable federal guardrails, according to a report by the New York Post citing his remarks on the program "Saturday in America."

Flock Safety is a private company that manufactures automated license plate readers and associated AI-analysis software. Its cameras are used by more than 5,000 law enforcement agencies and private communities across the United States, according to the company's own published figures. The devices capture vehicle images, license plates, and associated metadata, which are stored and made searchable across agency networks.

Beall's remarks did not target Flock Safety as a company specifically but focused on the policy gap: no federal statute currently defines retention limits, access controls, or audit requirements for AI-assisted surveillance tools used by local law enforcement agencies receiving federal funding. The relevant statutory landscape is governed primarily by the Fourth Amendment and patchwork state laws, with no unified federal framework in place as of the date of this report.

The absence of a federal standard has practical consequences for communities where Flock cameras have been installed through federal grant programs. The Edward Byrne Memorial Justice Assistance Grant program, administered by the Department of Justice Bureau of Justice Assistance, is one documented funding channel through which local agencies have purchased surveillance hardware. Grant award data is publicly available through USASpending.gov, though individual purchases of Flock equipment within those grants are not always itemized at the line level.

Congress has not passed legislation specifically governing AI-enabled license plate reader networks. The American Data Privacy and Protection Act, which passed the House Energy and Commerce Committee in July 2022 by a 53-2 vote according to the House committee record, stalled before a full floor vote and was not enacted. No equivalent bill has cleared both chambers in the current Congress.

The Federal Trade Commission has authority under Section 5 of the FTC Act to regulate unfair or deceptive data practices, but that authority has not been applied by rule or enforcement action specifically to law enforcement-adjacent surveillance vendors as of September 2026. The FTC's rulemaking calendar, publicly available on its website, does not list a pending rule on license plate reader data as of this writing.

At the state level, at least 15 states have enacted some form of license plate reader regulation, according to the National Conference of State Legislatures' tracked legislation database. Those laws vary in scope: some limit retention periods to 30 days, others to several years, and a minority impose no retention limit. What constitutes lawful query access also differs materially by state.

Beall's public remarks add a credentialed voice to a policy debate that has been active in academic and civil liberties circles for several years. The American Civil Liberties Union published a 2021 report documenting retention practices across more than two dozen agencies that use automated plate readers. That report found retention periods ranging from 48 hours to indefinite, with no uniform auditing standard across the agencies surveyed.

The Department of Homeland Security's Science and Technology Directorate has funded research into license plate reader standards, but no binding regulatory output from that research has been published in the Federal Register as of September 2026.

What remains unknown is whether any current or pending federal legislation would address the specific data-sharing architecture Beall described in his remarks. The full transcript of his September 5 appearance has not been released as a public document. A complete accounting of federal grant dollars that have funded Flock Safety deployments would require a targeted review of Byrne JAG subaward data on USASpending.gov cross-referenced with individual agency procurement records, neither of which has been compiled and published by a federal oversight body as of this report.