Nurse at Tennessee hospital steals fentanyl and AI misses it

A nurse at Erlanger Baroness, Tennessee’s largest hospital in Chattanooga, diverted and abused fentanyl left over from surgeries for months before being caught. The nurse exhibited signs of impairment while on duty, failed a drug test, and later admitted to stealing fentanyl daily, according to a Tennessee Board of Nursing consent order. The case is notable because Erlanger uses Sentri7, an AI-powered medication-monitoring software designed to flag missing drugs faster than humans. However, the system failed to detect inconsistencies in drug usage for months, despite fentanyl theft being a common and easily identifiable form of drug diversion. The Tennessee Board of Nursing’s order states that Sentri7 overlooked discrepancies that should have triggered alerts, exposing a potential failure in AI-driven security systems used in hundreds of U.S. hospitals. There is no legal requirement for hospitals to disclose AI software use or report malfunctions, leaving gaps in transparency and oversight. Experts, including Johns Hopkins Medicine’s David Rastall, criticized the lack of transparency in AI healthcare tools, warning that errors could go unaddressed and repeated. Jacob Smith, a pharmacist at Johns Hopkins, questioned how the software could miss such a basic issue, while Terri Vidals of Rxpert Solutions suggested user error might have contributed. The Drug Enforcement Administration requires hospitals to report lost or stolen drugs confidentially, but these reports do not include details about AI software involvement. This case marks the first publicly documented failure of AI drug diversion software, highlighting broader concerns about accountability in healthcare technology.