An artificial intelligence drug monitoring system marketed to hundreds of U.S. hospitals has failed a critical real-world test, according to state investigation records. The Sentri7 software, specifically designed to detect controlled substance diversion patterns, did not flag systematic fentanyl theft by a healthcare worker at a Tennessee medical facility over several months in 2025.
The failure raises urgent questions about the effectiveness and validation of AI-powered surveillance systems deployed across American healthcare infrastructure. Sentri7, developed by FDA-regulated technology companies, is supposed to analyze medication dispensing patterns and identify anomalies indicative of theft or misuse. Yet the system missed clear evidence of drug diversion at an operational hospital facility.
Experts and healthcare administrators are now scrutinizing the gap between AI system marketing claims and demonstrated performance in preventing controlled substance diversion—a persistent challenge affecting patient safety and institutional integrity.
Read the full article on GMJ Newsroom.
Was this article helpful?
