Glossary term
Glossary term
Governance and Compliance
A formal statement of management intent for responsible AI. It defines principles, scope, responsibilities, acceptable use, risk thresholds, documentation expectations, and how AI governance aligns with existing security, privacy, and compliance programs. The policy should be specific enough to drive exceptions, audits, procurement decisions, model release gates, and escalation actions.
Microsoft's Responsible AI Standard v2 (2022) codifies six principles into impact assessment requirements, with mandatory review by the Office of Responsible AI for sensitive uses.
Salesforce's Acceptable Use Policy for Einstein GPT explicitly prohibits use for legal advice, medical diagnoses, and political campaigning, enforceable through contract.
JPMorgan Chase introduced a policy in 2023 restricting employee use of public ChatGPT pending internal review, later replaced by an internal LLM Suite tool with documented use boundaries.