The Secret Rollout: How Flock Safety Landed in Pawtucket
In early 2021, Flock Safety offered a free 60‑day pilot of its automated license‑plate‑reader (ALPR) system to several Rhode Island municipalities, including Pawtucket, Cranston, and Woonsocket. The pilot installed 17 compact, solar‑powered cameras on utility poles, each angled to capture traffic flow on the busiest corridors.
By May 2023, local reporting suggested that Pawtucket would not move forward with a permanent deployment. Yet, in September 2023 the city reversed course, approving a two‑year contract worth $128,000 for a 20‑camera network covering Main Street, Broadway, Newport Avenue, Smithfield Avenue, and West Avenue. The decision was made without a public hearing, a formal board vote, or any community outreach—an omission that set the stage for the whistleblower episode.
Noel Pichardo’s Break‑the‑Silence Moment
Officer Noel Pichardo, a 2018 recruit who studied criminal justice at Salve Regina University, stumbled upon the hidden network during routine patrols. Noticing unfamiliar camera housings, he dug into internal memos and discovered that the department had already ordered the hardware, contradicting the earlier public statements.
Pichardo’s internal concerns quickly escalated into a public critique when he approached Valley Breeze, the local newspaper. In the October 11, 2023 article, he warned that the system represented a “grave threat” to privacy and gave him “1984 vibes.” He also reminded his colleagues, “You should never sell your integrity.”
The fallout was swift. On October 25, 2023, the Pawtucket Police Department opened an internal investigation into Pichardo’s conduct. The investigation, coupled with mounting pressure from the ACLU of Rhode Island, effectively ended his law‑enforcement career. His story illustrates the personal cost of confronting opaque surveillance programs.
Technical Deep Dive: Inside the Flock Safety Platform
Core ALPR Engine
Flock’s hardware combines a high‑resolution camera with edge‑processing chips capable of reading license plates in real time. The system uses optical character recognition (OCR) tuned for North American plate formats, achieving read rates above 95 % under optimal lighting.
Advanced Visual Fingerprinting
Beyond plates, Flock advertises “advanced recognition” that can flag vehicles by bumper stickers, roof racks, and other distinguishing features. This capability relies on convolutional neural networks trained on millions of vehicle images. While technically impressive, it raises the specter of mass vehicle profiling, a concern echoed in privacy‑focused AI debates such as those covered in the YouTube AI policy article: https://ltdeveloperblogs.github.io/posts/youtube-clarifies-policies-around-ai-slop-and-upsetting-videos.
Movement Reconstruction Across Jurisdictions
One of the platform’s headline features is the ability to reconstruct a vehicle’s path camera‑by‑camera, even across municipal boundaries. Data is stored centrally and made searchable by any participating agency. In practice, a single license‑plate query can return billions of sightings per month, enabling investigators to trace a car’s route over hundreds of miles.
Inter‑Agency Data Sharing
Flock markets its network as “the nation’s largest fixed license‑plate‑reader network.” By default, records are shared with partner law‑enforcement agencies, meaning a query entered in Boston could surface data captured in Pawtucket. This openness, while useful for multi‑jurisdictional investigations, also creates a single point of failure for data security—a risk highlighted by recent high‑profile exploits like the Zoom zero‑day vulnerability: https://ltdeveloperblogs.github.io/posts/zoom-flaw-let-an-attacker-take-over-your-device-including-iphone-and-mac.
Security Posture
The cameras are solar‑powered and physically hardened, but the backend relies on cloud‑based APIs and a web portal for data access. Penetration testing reports from independent firms have identified potential API authentication flaws, reminiscent of the Zoom annotation flaw that allowed unauthorized data extraction: https://ltdeveloperblogs.github.io/posts/zoomsday-hack-uncovered-using-fewer-than-20-ai-prompts. If exploited, an attacker could retrieve raw plate‑sightings, effectively turning a public safety tool into a surveillance trove.
Legal, Ethical, and Privacy Implications
Constitutional Concerns
The Fourth Amendment protects against unreasonable searches. Courts have been split on whether continuous, automated plate collection constitutes a search. In Carpenter v. United States (2018), the Supreme Court recognized that digital location data warrants protection. ALPR data, especially when retained for extended periods, may fall under the same scrutiny.
Transparency and Public Oversight
Rhode Island’s Open Records Law requires municipalities to disclose surveillance contracts and data retention policies. The Pawtucket deployment violated both the spirit and, arguably, the letter of the law by proceeding without a public hearing. The ACLU’s criticism underscores a broader trend: municipalities often adopt “surveillance as a service” models without clear governance frameworks.
Data Retention and Sharing
Flock’s default retention period can extend to 90 days or more, depending on the contract. Longer retention amplifies privacy risks, especially when data is shared across state lines. Whistleblower protections in Rhode Island are limited for law‑enforcement personnel, leaving officers like Pichardo vulnerable when they expose policy breaches.
Impact on Community Trust
Surveillance programs can erode trust between police and the communities they serve.
Read the full breakdown originally published at https://ltdeveloperblogs.github.io/posts/the-cop-who-took-on-flock/
Top comments (0)