Skip to content
AAOE
Draft3 min read

Human-Governed Engineering

Purpose

This Standard defines Human-Governed Engineering as the governing principle for all engineering activities involving artificial intelligence.

Engineering decisions remain under human governance regardless of the degree of AI assistance.

Standard

Engineering teams shall ensure that governance of engineering work remains a human responsibility.

AI systems may generate, analyze, review, or recommend engineering artifacts.

Humans shall retain responsibility for engineering approval, architectural decisions, governance, and accountability.

Authority may be delegated to AI for execution, but never for governance.

Principles

Human-Governed Engineering shall:

  • preserve human accountability
  • maintain engineering governance
  • improve engineering quality
  • improve decision traceability
  • encourage effective AI collaboration
  • preserve organizational responsibility
  • enable responsible engineering automation

Responsibilities

Humans are responsible for:

  • Engineering Governance
  • Architecture Decisions
  • Standard Approval
  • Architecture Reviews
  • Risk Acceptance
  • Compliance Decisions
  • Final Approval

AI systems may assist with:

  • Specification Drafting
  • Design Proposals
  • Code Generation
  • Documentation
  • Testing
  • Analysis
  • Implementation Support

AI recommendations remain subject to human review.

Engineering Workflow

The preferred engineering sequence is:

Engineering Intent
        ↓
AI-Assisted Draft
        ↓
Human Review
        ↓
Architecture Review
        ↓
Approval
        ↓
Implementation
        ↓
Validation

Governance decisions shall require human approval.

Traceability

Governed engineering should remain traceable to:

  • Specifications
  • Standards
  • Architecture Reviews
  • Implementation Reports
  • Decision Records

Human approvals should remain auditable.

Exceptions

Routine automation may execute previously approved engineering work.

Governance decisions shall not be fully delegated to autonomous systems.

Any exception shall be explicitly documented and approved.

Relationship to ADR-0004

This Standard operationalizes ADR-0004 — Human Governance of AI Contributions, the accepted decision establishing that AI agents may propose, draft, and prepare engineering work, while a human maintainer reviews and approves decisions before they become accepted, merged, released, or published as AAOE.

This Standard defines how engineering teams apply that principle in day-to-day engineering work.

Relationship to the Operating Model

The Operating Model applies this Standard through:

  • Work Orders
  • Architecture Reviews
  • Implementation Reports
  • Human Approval Gates
  • Engineering Governance

Compliance

Engineering work conforms to this Standard when:

  • Human governance is preserved.
  • Engineering decisions remain accountable.
  • AI-generated artifacts are reviewable.
  • Human approvals are documented.
  • Governance decisions remain traceable.

Future Evolution

Future Standards may define:

  • Multi-Agent Governance
  • Autonomous Engineering Boundaries
  • AI Review Policies
  • Engineering Approval Workflows
  • Governance Metrics
  • Agent Accountability Models