Evaluating both AI algorithms and human users is key for safe adoption in high-stakes healthcare settings, according to an Ohio State study.
Key Details
- 1Researchers studied 462 nursing students and professionals using AI-assisted patient monitoring simulations.
- 2Accurate AI predictions improved decision making by up to 60%, but inaccurate AI caused over a 100% drop in correct decisions.
- 3Explanations and supporting data had minimal impact; participants were highly influenced by AI predictions.
- 4The study calls for simultaneous evaluation of both algorithms and clinical users in safety-critical settings.
- 5Findings appear in npj Digital Medicine (DOI: 10.1038/s41746-025-01784-y).
Why It Matters
This work highlights the risks of over-reliance on AI in clinical workflows and underscores the necessity for robust, joint evaluation protocols in radiology and other safety-critical healthcare applications, ensuring that human-AI teams can handle both good and poor system performance.
Source
EurekAlert
Related News
•EurekAlert
New Framework Compares AI Segmentation Without Ground Truth Annotations
Researchers introduce an open-source approach for evaluating AI anatomy segmentation models in medical imaging without requiring ground truth annotations.
•EurekAlert
FDA Approves Johns Hopkins AI Tool for Early Sepsis Detection
FDA clears an AI-driven system developed by Johns Hopkins to detect sepsis up to 48 hours earlier and reduce mortality rates.
•EurekAlert
AI-Driven Handheld Endomicroscope Enhances Early Cancer Detection
Researchers develop PrecisionView, a handheld AI-powered endomicroscope for real-time, high-resolution cancer diagnostics.