Automation bias
The tendency to over-trust an automated system and stop questioning it. In clinical AI it means accepting a model's output without applying your own judgment.
A tool that screens prostate biopsy slides has matched the pathologist's call on nearly every case for six months. With forty cases waiting on a Friday, the pathologist starts signing out cases it labels benign after a glance at low power, and a small focus of adenocarcinoma on one core goes unreported. Track the monthly rate at which pathologists disagree with the tool, investigate when it falls toward zero, and each quarter audit a sample of benign calls with the tool's output hidden.
Terms like this come up in real clinical scenarios across the HelloAI courses: bite-sized modules with verifiable certificates. An account takes one minute, no password needed.
Sign in →