RESPONSIBLE AI, COMPLIANCE, AND RISK

ARTIFICIAL INTELLIGENCE ASSISTANT PROMPTS [START OF PROMPT] Classify this new agent feature's risk level (low/moderate/high) using a risk-management framework approach (such as the NIST AI Risk Management Framework) and specify exactly what review process that risk level requires before launch: feature = [FEATURE DESCRIPTION, e.g. "agent auto-recommends candidates for a job posting"]. Output format: risk classification with justification, plus the required review/approval steps for that risk level. [END OF PROMPT] [START OF PROMPT] Design a bias-testing plan for this screening/matching agent and run it conceptually against this sample output set: agent = [AGENT DESCRIPTION, e.g. Resume Review Agent], sample outputs = [PASTE LAB MATCH RESULTS / RANKED CANDIDATE LIST HERE]. Specify at least three concrete tests for unfair treatment across protected characteristics, apply them to the sample data, and state what disparity threshold should block a launch. Output format: test plan table plus an applied-results summary flagging any concern found. [END OF PROMPT]