Updated
Updated · CBS New York · Jun 1
Sentri7 Missed 5 Fentanyl Diversions at Erlanger Hospital, Exposing AI Oversight Gaps
Updated
Updated · CBS New York · Jun 1

Sentri7 Missed 5 Fentanyl Diversions at Erlanger Hospital, Exposing AI Oversight Gaps

3 articles · Updated · CBS New York · Jun 1
  • Tennessee nursing records show Sentri7 failed for about 4 months to flag a nurse anesthetist’s fentanyl diversion at Erlanger Baroness, despite roughly five missed incidents and other dispensing inconsistencies.
  • John Stevenson admitted taking leftover fentanyl from March to June 2025—escalating to daily use—after co-workers noticed slurred speech and nodding off; he was fired, and his license was put on probation.
  • Erlanger and Wolters Kluwer declined to explain what happened, while the board order said the software was in an “initial learning phase,” a claim a company executive disputed by saying Sentri7 is trained on 9 to 12 months of historical data.
  • The case is a rare public look at a possible failure in AI anti-diversion tools used with little disclosure at U.S. hospitals, which are not required to report software malfunctions even as experts say diversion is widespread.
A hospital’s AI watchdog missed months of opioid theft. Who is ultimately held accountable for the silent failure?
AI's secret failures are buried by hospitals. How can patients trust technology that operates without public oversight?

When AI Falls Short: The Erlanger Fentanyl Diversion Incident and the Future of Drug Safety Oversight in Healthcare

Overview

In 2025, the Erlanger Hospital fentanyl diversion incident drew intense attention to the effectiveness of drug diversion protocols and the transparency of healthcare institutions. The hospital faced scrutiny over its handling of the situation and its use of advanced detection technology, particularly after preparing but never releasing a statement about its AI-powered Sentri7 system. This silence, along with the hospital's decision to decline further comment, left the public with little information about how the diversion occurred or was detected. As a result, the incident raised broader concerns about human oversight, safeguards, and trust in both technology and institutional transparency.

...