act-responsible-ai-review

name	act-responsible-ai-review
description	Reviews an AI-related decision for stakeholder impact, privacy, responsibility, transparency, contestability, and reversibility.

1. Describe the intended outcome, affected stakeholders, context of use, and the decision or action the system can influence.
2. Identify foreseeable benefits and harms, including unequal impact, exclusion, privacy exposure, autonomy, accessibility, misuse, and error consequences.
3. Assess data provenance, minimization, consent or lawful basis where relevant, retention, access controls, and whether sensitive inferences exceed the stated purpose.
4. Define meaningful human responsibility: who can understand, challenge, override, remediate, and monitor consequential outcomes. Match transparency and contestability to the impact on people.
5. Compare alternatives, including not using AI, and recommend safeguards, evidence, approval, or a pause where uncertainty or harm is material. Do not substitute this review for legal, policy, or domain-expert advice.