Public Sector and Academia
AI governance patterns from NSW, Yale, US DOJ, and government playbooks.
NSW (New South Wales, Australia)
NSW operates one of the most clearly documented public-sector AI governance models.
Structure
- AI Review Committee (AIRC). A central body that reviews only high-risk and critical-risk AI systems. Referrals are coordinated through the NSW Office for AI.
- Agency accountability. Agencies retain final accountability for their AI systems. The AIRC provides expert advice, not mandates.
- Risk-based referral. The AI Assessment Framework automatically identifies which systems must be referred to the AIRC. Low-risk systems move without central review.
Key Practices
- Mandatory self-assessment. The AI Assessment Framework is mandatory for all NSW Government AI use under Circular DCS-2024-04. It replaced earlier, more subjective assessments with a faster, standards-aligned approach that NSW reports can reduce assessment time from days to under 30 minutes for low-risk systems.
- Five ethics principles. The framework is structured around community benefit, fairness, privacy and security, transparency, and accountability.
- Assurance integration. AI assurance is embedded in the broader NSW Digital Assurance Framework, with projects exceeding $5 million receiving additional central oversight.
- Standards alignment. The framework is aligned with the Commonwealth National Framework for the Assurance of AI in Government and references the EU AI Act.
Lesson for Your Council
NSW demonstrates the referral model: the central body reviews only what truly needs central review. Everything else is handled closer to the team. This prevents bottlenecks and keeps the council focused on high-impact cases. The mandatory-but-lightweight self-assessment is a pattern worth borrowing.
Yale University
Yale's AI governance reflects the reality of a large, decentralized institution. Rather than a single AI council, Yale operates through multiple overlapping governance structures, each authoritative in its own domain.
Structure
- Yale Task Force on Artificial Intelligence (YTAI). Assembled by Provost Scott Strobel in January 2024 to review AI activity through a university-wide lens. The Task Force published its report in June 2024, establishing strategic direction for AI use across teaching, research, and operations.
- IT Governance Committees. Yale's Academic IT Steering Committee and Research IT Steering Committee provide oversight for systems supporting educational and research missions respectively, including AI tools.
- Digital/AI Implementation Advisory Committee (DAIAD). Performs comprehensive review of AI assessments and provides lifecycle oversight of AI solutions, ensuring vetting and risk evaluation before deployment.
- Healthcare AI governance. The Joint Health Data Governance Council and Yale New Haven Health System's Enterprise Healthcare AI Governance process manage clinical AI tools through a separate, domain-specific review process.
- Research oversight. Institutional Review Boards govern AI use involving human subjects research.
Key Practices
- Provost-level sponsorship. The YTAI was commissioned directly by the Provost, giving AI governance executive-level visibility and authority.
- Domain-specific authority. Healthcare AI, research AI, and administrative AI each have their own governance paths. There is no single committee that reviews everything.
- Centralized guidance. The AI at Yale site provides university-wide guidance, tool recommendations, and an AI review framework for system selection, creating a shared foundation even though governance is distributed.
Lesson for Your Council
Yale shows what federated governance looks like in practice at a large, diverse institution. The key insight is that multiple governance bodies can coexist effectively when each has a clear domain, a shared set of principles, and a connection to executive leadership. Organizations with distinct business units or regulated domains (healthcare, research, finance) should consider this pattern.
US Department of Justice
The DOJ provides one of the clearest examples of inventory-as-governance in the US federal government.
Structure
The DOJ AI governance operates within the framework established by Executive Order 13960 (Promoting the Use of Trustworthy Artificial Intelligence in the Federal Government, December 2020) and OMB Memorandum M-25-21 (Accelerating Federal Use of AI through Innovation, Governance, and Public Trust, April 2025).
Key Practices
- Annual AI inventory. The DOJ inventories all AI use cases annually, submits the inventory to OMB, and publishes releasable entries on its website. The 2025 consolidated federal inventory is published on GitHub.
- Public transparency. Disclosed government AI use across all federal agencies increased by 70% in 2025, demonstrating that inventory discipline drives visibility.
- Governance framing. The inventory is explicitly tied to advancing governance, accelerating innovation, and ensuring public trust.
Lesson for Your Council
The DOJ example shows that inventory discipline is not just compliance. It is a governance practice that builds internal and external trust. Publishing your AI inventory (even internally) forces accountability and surfaces systems that might otherwise operate without oversight.
Government Playbooks
Several governments have published AI playbooks that serve as useful references for council design:
| Country | Resource | Focus |
|---|---|---|
| UK | AI Playbook for the UK Government | Public-sector AI implementation guidance |
| Singapore | Responsible AI Playbook + AI Verify | Principles and testing framework |
| Canada | Algorithmic Impact Assessment | Mandatory risk assessment for automated decisions |
| Australia (NSW) | AI Assessment Framework | Risk self-assessment for agencies |
| Australia (Queensland) | FAIRA | Transparency and risk identification |
These playbooks are especially useful as input to your council's intake and review templates.
Sources
- NSW AI Review Committee: digital.nsw.gov.au/policy/artificial-intelligence/ai-governance-assurance-and-frameworks/ai-review-committee
- NSW AI Assessment Framework: digital.nsw.gov.au/policy/artificial-intelligence/ai-governance-assurance-and-frameworks/nsw-ai-assessment-framework
- Yale Task Force on Artificial Intelligence: provost.yale.edu/yale-task-force-artificial-intelligence
- Report of the Yale Task Force on Artificial Intelligence: ai.yale.edu/report-of-the-yale-task-force-on-artificial-intelligence
- Yale AI Guidance: ai.yale.edu/guidance
- Yale IT Governance Committees: its.yale.edu/about-it/it-committees/governance-committees
- DOJ AI Inventory: justice.gov/ai/ai-inventory
- 2025 Federal Agency AI Use Case Inventory (GitHub): github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory
- FedScoop, "Disclosed government AI use increased by 70% in 2025": fedscoop.com/disclosed-government-ai-use-increased-in-2025-omb/