Creepy Cops Track Women, Audits Explode

Young woman speaking with two officers at a desk — Photo: Ground Picture / Shutterstock

When law-enforcement databases become dating tools, the problem isn’t a few bad actors; it’s an oversight architecture that lets personal curiosity masquerade as official purpose until an audit trips the wire.

At a Glance

Pinellas County investigators say two former deputies repeatedly queried restricted systems to look up women for personal reasons, triggering criminal charges and agency reforms.
One deputy allegedly ran a 17-year-old and her family across multiple databases; another logged hundreds of queries on a woman he hoped to date, including via automated license plate readers (ALPRs).
The case illustrates a common failure mode in policing technology: authorized access used for private aims, detectable only through audit logs, reason codes, and supervisory review.
Sensible controls for ALPRs and criminal-justice data exist—purpose coding, immutable logs, periodic audits—but must be enforced to work.

What the Pinellas cases show about database power and human temptation

The Pinellas County Sheriff’s Office arrested two of its own after internal probes concluded the deputies mined restricted systems for personal reasons, not legitimate police work. According to the agency, former deputy Travis Stanton used patrol-car access to run a 17-year-old and her family through multiple databases; he was booked on one count of accessing an electronic device without authority and one count of official misconduct. Former deputy Karl Gwynne, in a separate investigation sparked by an audit trail, allegedly queried a woman’s records 381 times across five systems between January 2025 and July 2026, including through the agency’s ALPR platform. Local reporting adds that more than half of those queries occurred when Gwynne was off duty, a detail consistent with private use rather than policing need.

Neither deputy’s searches, the sheriff said, tied back to a case file or investigatory purpose. Supervisors reported that Stanton at first keyed a “verifying an identity” rationale before conceding there was no official reason—he was “curious”—while Gwynne reportedly insisted on “official use” despite the audit pattern suggesting otherwise. The charges are allegations, and the courts will test them. But the mechanism by which misconduct is detected here—post hoc review of access logs and stated purposes—reflects how these systems are meant to catch insider misuse without impeding legitimate work.

How the technology works—and how it’s misused

Modern criminal-justice data flows through layered systems: state driver and vehicle records, in-state criminal justice networks, the national NCIC index, local RMS/CAD systems, and, increasingly, ALPR platforms that capture and index plate reads with time and location metadata. Access is credentialed by user ID, and well-designed deployments require a “purpose code” or case number for each query. Unlike external hacking, which breaches from the outside, these cases involve credentialed insiders who “piggyback” on legitimate privileges, leaving a forensic trail that—if reviewed—can distinguish on-duty policing from personal voyeurism. ALPRs intensify the risk: a single officer can, with a few keystrokes, assemble a longitudinal map of where a person’s car has been, converting an occasional lookup into a pattern-of-life dossier if no guardrails intercede.

That capacity is not theoretical. The audit that flagged Gwynne reportedly began with ALPR queries, then widened across systems as investigators followed the user’s footprint. The volume—hundreds of lookups concentrated on one subject—functioned as a statistical anomaly that no plausible caseload could justify. This is exactly what a competent internal audit should surface: outlier patterns, missing case ties, and off-duty spikes that defy operational need.

This is not a one-off: the base rate of insider misuse

Across the United States, insider misuse of law-enforcement databases is a recurring governance problem. The Associated Press documented hundreds of discipline cases over a three-year span for confidential database abuse—often looking up exes, neighbors, or romantic interests—underscoring that the failure mode is ordinary temptation amplified by powerful tools, not elaborate conspiracy. Minnesota’s 2013 legislative audit spurred changes after finding that more than half of 11,000 users made questionable queries in the state driver system, a scale that points to systemic, not idiosyncratic, risk. ALPR-specific policy reviews from academic and policy shops repeatedly land on the same prescription: instrument the access, force users to declare purpose, and audit the declarations against outcomes—with external oversight where possible.

It is telling that the Pinellas cases turned on log review. When oversight is real, misuse is discoverable. When oversight is lax, convenience and curiosity win. Agencies that treat audits as a compliance box-check rather than an operational function tend to learn the hard way—through scandal—that insider access is its own attack surface.

Law and policy: what “unauthorized” means for insiders

Courts have narrowed certain federal computer-crime theories for insiders—most notably in Van Buren, where the Supreme Court held that a police officer who had authorized access did not violate the Computer Fraud and Abuse Act merely by breaching use policy. But that does not legalize policy violations; it simply locates the misconduct in other statutes (state computer-crime laws, official misconduct) and in administrative discipline. In practice, agencies pair criminal charges where state law supports them with termination or decertification. The deterrent backbone is less the distant threat of prosecution than the near-certainty that improper queries will be detected, investigated, and career-ending.

That certainty depends on design. Purpose codes must be mandatory and meaningful, not rubber stamps. Audit logs must be immutable and attributed to individual users. Supervisors must receive exception reports—high-volume queries, subject-focused runs, off-duty access—on a schedule measured in days, not months. And ALPR vendors should deliver native audit tooling that surfaces anomalous use automatically, because plate data enables movement tracking in a way traditional records do not.

What competent oversight looks like in practice

Several concrete controls separate programs that work from those that drift. First, require a case number or incident ID for every sensitive query, with dropdown purpose codes that map to policy; free-text rationales are audit catnip for after-the-fact storytelling. Second, enforce least-privilege access—detectives do not need the same reach as traffic units, and off-duty access should default to locked. Third, operationalize auditing: build exception thresholds that auto-alert when a user queries the same subject repeatedly, performs lookups while off duty, or queries outside their assignment area. Fourth, couple training with consequences: years of policy PowerPoints matter little if every misuse is “counseled” rather than disciplined.

Why it matters—beyond these two arrests

Trust in policing is cumulative. Each time an officer weaponizes privileged access for private aims—whether to pursue a teenager or track a potential date—it reinforces public suspicion that surveillance tools can be turned inward. The Pinellas cases are, paradoxically, a backhanded proof that oversight can work: audits detected anomalies; leadership moved; charges followed. The lesson is not to abandon tools like ALPRs or criminal-justice databases—both solve real crimes daily—but to govern them as if temptation were a constant, because it is. Programs that make improper queries hard to perform and easy to detect will keep both the technology and the badge worthy of the authority they carry.