Healthcare AI Glossary

A

AI Governance (Healthcare): The standing process a healthcare organization uses to decide which AI tools it approves, on what evidence, who owns each tool after deployment, and what triggers turning it off.
Algorithmic Bias (Healthcare): Systematic difference in an AI model's performance across patient groups, usually introduced through unrepresentative training data or a target variable that stands in poorly for the thing being predicted.
Ambient AI Scribe: An AI tool that passively listens to patient-physician conversations and automatically generates clinical documentation such as SOAP notes, without requiring manual dictation.

Clinical Decision Support (CDS): Software tools that provide clinicians with knowledge and patient-specific information to enhance clinical decision-making at the point of care.

Generative AI in Healthcare: The application of large language models and other generative AI systems to healthcare tasks including clinical documentation, patient communication, medical education, and research synthesis.

Model Drift: Degradation in a deployed model's performance over time caused by changes in the population, the data, or clinical practice, even when the model itself has not been modified.

Predictive Analytics (Healthcare): The use of statistical or machine learning models on clinical and operational data to estimate the likelihood of a future event, such as deterioration, sepsis onset, readmission, or a missed appointment.

Responsible AI (Healthcare): A set of commitments about how healthcare AI is built and deployed, most useful when reduced to testable checks: representative validation, actionable explainability, named accountability, and ongoing performance measurement.

Software as a Medical Device (SaMD): Software intended to perform a medical purpose, such as diagnosis, treatment, prevention or prediction, without being part of a hardware medical device, which brings it inside FDA device authorities.