Free preview: AI Governance Policy

Policy Statement

[Company Name] permits the use of AI tools and AI-enabled features only when the specific use has been approved through the workflow defined in this policy, recorded in the AI system inventory, and assessed for risk in accordance with the AI Risk Assessment Procedure. AI must support, not replace, accountable human judgment: a named individual remains responsible for every business decision and every work product, regardless of whether AI contributed to it.

This policy is the parent document of the organization's AI governance documentation set. Day-to-day rules for individual users are defined in the AI Acceptable Use Policy; technical and procedural detail is delegated to the procedures and standards listed in the Related Documents section. Where this policy conflicts with a subordinate document, this policy prevails.

Governance Principles

All AI adoption decisions at [Company Name] must be made consistent with the following principles. Approvers must be able to demonstrate, in writing, how a proposed AI use satisfies each principle.

Roles and Responsibilities

The following roles are mandatory. One person may hold more than one role in a small organization, except that the [Role, e.g. Owner or Managing Partner] may not also act as the [Role, e.g. AI Governance Lead] for the same approval decision. Role assignments must be documented by name and reviewed at least annually by the [Role, e.g. Owner or Managing Partner].

Approval Workflow for New AI Uses

No new AI tool, AI-enabled feature, or materially new use of an already-approved tool may be placed into business use before completing the following workflow. A 'materially new use' includes a new data category, a new user population, a new decision the tool influences, or use of a previously disabled AI feature in existing software.