Cross-Cutting Work Groups
Description
Develop a Risk Categorization Tool that supports early AI governance by helping identify Low, Medium, or High risk levels across several risk modifiers related to prioritized risk domains.
Meeting Cadence: Thursdays (biweekly) from 2-3pm ET
Timeline
Q2 2025 - Q4 2026
Timeline:
Q4 2025: develop Risk Categorization Tool (v1); includes risk domain of (1) life & patient safety
Q1 2026: develop Risk Categorization Tool (v2); includes risk domains of (1) life & patient safety and (2) technology & data
Q2 2026: stress-test Risk Categorization Tool (v2) with work group members
Tasking:
The Risk Work Group’s scope of work falls within four primary phases:
Risk Categorization (active)
Risk Assessment (future)
Risk Mitigation (future)
Risk Monitoring (future)
Output:
CHAI’s Risk Categorization Tool (v2). This tool was developed by the CHAI Risk Work Group and is for health systems (any size) and teams responsible for pre-deployment risk review. This tool supports early AI governance by helping identify Low, Medium, or High risk levels across several risk modifiers related to the risk domains: (1) Life & Patient Safety and (2) Technology & Data. The tool highlights key dimensions of risk to guide further assessment, mitigation, and monitoring.
Supplementally, the CHAI Risk Work Group developed a detailed information mapping between the CHAI Risk Categorization Tool (v2), NIST AI Risk Management Framework (RMF), and EU AI Act.
