Jason Killinger a long haul truck driver from Nevada spent eleven hours in police custody after facial recognition software at the Peppermill Casino in Reno wrongly flagged him as a one hundred percent match to a banned patron. The incident unfolded on September 17, 2023, while he was attempting to leave following a winning streak at the craps tables. Despite presenting multiple forms of valid identification the responding officer arrested him solely on the basis of the software alert prompting a federal lawsuit now seeking damages from both the officer and the city for lost wages emotional distress and the infringement on his freedom.
Body camera footage captures Killinger already in security custody when Officer Richard Jager arrived at the scene. He calmly but firmly protested the error offered his Nevada drivers license a UPS paystub his vehicle registration a debit card and his Teamsters union card and explained he had no connection to the banned individual Michael Ellis. The officer acknowledged discrepancies in age weight eye color and signatures yet relied on the alert to arrest him as a John Doe for fingerprint verification at the county jail where he was cleared later that afternoon.
National Institute of Standards and Technology evaluations of facial recognition technology reveal that while leading algorithms can achieve false positive identification rates as low as zero point three percent in controlled high quality mugshot comparisons real world performance drops significantly with casino surveillance cameras lower resolution images and varying angles. NIST FRVT reports from 2023 through 2026 show that unconstrained environments such as gaming floors introduce error rates that can climb to nine percent or higher and demographic differentials persist with false positives occurring up to one hundred times more frequently for certain racial and gender groups compared to others. This case exemplifies how a claimed one hundred percent match can still result in wrongful detention when human oversight and corroborating evidence are absent.
Peppermill Casino reached an undisclosed settlement with Killinger shortly after the incident yet has not publicly announced any overhaul to its facial recognition practices. Its current privacy policy explicitly states that biometric information including facial recognition data collected through cameras in public areas and entrances may be used for safety security and surveillance purposes. The policy confirms ongoing collection and storage of such data without indicating limits on sole reliance for identifying patrons or requirements for additional verification steps before involving law enforcement.
Reno Police Department records show that no formal training or policy on facial recognition existed at the time of the arrest and only rudimentary sessions were introduced in 2025. As of April 2026 the department still lacks a requirement that officers treat software matches as advisory rather than sufficient probable cause. Officer Jager later testified that he had conducted hundreds of similar arrests based on the technology alone and that proper training would have prevented this outcome.
This situation underscores the tension between technological efficiency and constitutional protections in everyday public spaces. When systems touted for near perfect accuracy in laboratory settings produce errors in practice the consequences fall hardest on ordinary individuals going about their routines. Mandating layered verification protocols and regular independent audits could bridge the gap between vendor claims and real world reliability while preserving the benefits of the tools.
