AI tool detects delirium severity in ICU patients

An AI tool analyzes facial expressions and movement to detect delirium severity in ICU patients, aiming for earlier diagnosis and better clinical decisions. The technology, developed by a transdisciplinary team, provides passive, real-time feedback to clinicians. An accompanying ethics roadmap guides responsible AI integration in critical care. The adaptive ethics framework emphasizes transparency, accountability, and patient-centeredness, developed with input from ICU survivors, caregivers, and clinicians.