Back to News
SecurityAI Understanding briefing

University of Michigan police contract for AI licence‑plate cameras never installed

A FOIA request shows the University of Michigan Police Department paid $127,500 for 15 Flock Safety AI cameras that were never deployed, sparking privacy concerns on campus.

5 min readRead the linked source
Source-provided image accompanying University of Michigan police contract for AI licence‑plate cameras never installed

What happened

The University of Michigan Police Department (UMPD) signed a $127,500 contract with surveillance‑technology firm Flock Safety in February 2025 for 15 Falcon Flex licence‑plate‑reader cameras, associated software and battery packs. Records obtained through a Freedom of Information Act request, reported by The Michigan Daily and cited by The Times of India, reveal that none of the cameras were ever installed or activated. UMPD confirmed the equipment remains unused, and it is unclear whether any payments were ever made. The contract allowed the cameras to capture vehicle data, retain it for up to 30 days (later reduced to seven days by Flock), and share footage with law‑enforcement agencies without a warrant under a “good‑faith” clause. The agreement also granted reciprocal access to other agencies in Flock’s Community Network, potentially expanding data sharing beyond the university.

According to a FOIA request filed by a student group, the University of Michigan Police Department entered into a contract on 3 February 2025 with Flock Safety, a company that markets AI‑powered automated licence‑plate‑reader (ALPR) systems. The agreement stipulated an annual payment of $63,750 for a two‑year term, covering 15 Falcon Flex cameras, the FlockOS Essentials software platform, Vehicle Fingerprint AI technology, and battery packs.

The Michigan Daily, which first reported on the contract, noted that the agreement was not publicly announced at the time. When asked for comment, UMPD deputy chief Melissa Overton confirmed that the cameras had never been installed or activated, but did not explain why the university chose not to proceed with deployment or whether any portion of the contract was ever paid.

The contract’s language allowed the cameras to capture a vehicle’s registration number and a suite of visual attributes, then automatically delete the data after 30 days unless manually archived. Flock Safety later shortened its default retention period to seven days. Section 5.3 of the agreement permitted the company to disclose footage to law‑enforcement agencies, government officials or third parties without a warrant if it believed such disclosure was legally necessary.

Beyond the campus, the contract granted UMPD access to Flock’s Community Network—a nationwide collection of privately owned cameras—enabling statewide or nationwide vehicle searches. Conversely, the agreement allowed reciprocal access for other participating agencies to search data captured by the university’s 15 cameras, should they ever be installed.

Source details: timesofindia.indiatimes.com

Why it matters

The episode highlights how AI‑enabled surveillance tools are being procured by public‑sector institutions without transparent oversight. Even though the cameras were never deployed, the contract’s terms illustrate the breadth of data that could be collected—vehicle make, model, colour, bumper stickers and damage—allowing detailed tracking of individual movements over time. The ability to share footage without a warrant raises significant civil‑liberties questions, especially on a campus where students and faculty may be monitored without suspicion. The deal also coincided with a broader expansion of campus cameras and a revision of university policies that removed oversight committees, intensifying concerns about unchecked surveillance. For policymakers, the case underscores the need for clear regulations on automated licence‑plate readers, data retention limits, and third‑party access. For universities, it serves as a cautionary tale about procuring AI‑driven security tools without public disclosure or community input.

The deal illustrates how AI surveillance tools can be embedded in public‑sector contracts with minimal public scrutiny. Even though the equipment was never used, the contractual provisions reveal the potential for extensive vehicle tracking and data sharing that could affect anyone driving on or near campus.

Privacy advocates, including law professor Michael Steinberg and ACLU student leader Radha Tanner, argue that such technology shifts policing from reactive investigations to pervasive data collection on all campus occupants. The lack of a warrant‑based disclosure requirement raises constitutional concerns about unreasonable searches and the potential for mission creep beyond legitimate public‑safety purposes.

The timing of the contract aligns with a broader campus‑wide security upgrade that saw new cameras installed at building entrances and the launch of a Security Technology Enhancement Project. Simultaneously, the university removed explicit restrictions on surveillance camera use and dissolved a student‑faculty oversight committee, reducing institutional checks on how such technology is deployed.

At the state level, Michigan lawmakers from both parties have called for tighter regulation of ALPR systems, citing similar privacy worries. The University of Michigan case could become a focal point for legislative hearings or new statutes governing data retention, third‑party sharing, and transparency requirements for AI‑driven surveillance.

Interactive Mechanism

Interactive Mechanism: How It Actually Works

Explore the underlying technology behind this development interactively.

System Requirements:
Best ArchitecturePure RAGRecommended pattern
Hallucination RiskVery LowGrounding efficacy
Update Cost$0 (Vector sync)Ongoing maintenance
Core takeaway: Fine-tuning teaches models how to speak (form, style, syntax); RAG teaches models what to say (verifiable facts). Never use fine-tuning alone for factual memory.
Interactive Concept Check+10 Points
AI Ethics Quiz

Why can ethical evaluation not be reduced to one model score?

What to watch next

Future actions by the University of Michigan’s Board of Regents, the state legislature, and Flock Safety will shape how AI surveillance is governed on campuses. Watch for any legal challenges or policy proposals aimed at tightening data‑sharing provisions and requiring warrants for footage release. Monitor whether UMPD decides to install the cameras or cancel the contract entirely, and whether the university revises its procurement transparency rules. Finally, observe how other universities respond—whether they adopt similar AI licence‑plate readers or implement stricter privacy safeguards in light of this exposure.

Legislative activity: Michigan’s state legislature may introduce bills that limit warrantless data sharing and set stricter retention periods for ALPR footage. Tracking any bills or hearings that reference the University of Michigan contract will indicate how quickly policymakers respond to such disclosures.

University governance: The Board of Regents could be prompted to review procurement policies, reinstate an oversight committee, or require public reporting of AI‑related security contracts. Any formal policy changes or public statements from the university will signal a shift toward greater transparency.

Flock Safety’s response: The company may update its contract language, data‑retention policies, or public‑facing documentation in reaction to the scrutiny. Monitoring Flock’s website, press releases, or future FOIA filings will reveal whether it adjusts its practices.

Campus activism: Student groups may file lawsuits, demand audits, or organize protests demanding the removal of all AI surveillance equipment. The scale and outcomes of such activism could influence other institutions considering similar technology.

Related guides & quizzes

Found this useful?