DCO Policy Watch – Eighth Edition

March 2026

Welcome to the eighth edition of the DCO Policy Watch: Governing the Cognitive State – AI in Government.

This edition examines the rise of AI in government and the governance challenge it creates. Governments are moving beyond experimentation. Analytical AI is already used in areas such as fraud detection, social benefit triage, and infrastructure maintenance. Generative AI is reshaping citizen-facing services and internal productivity. The report’s core message is that governments must be able to govern AI at scale while preserving transparency, due process, and public trust.

Three pillars for AI in government

  • Delivering public value: Service transformation and automation: AI in government should be justified by measurable public value, including faster decisions, fewer errors, better access, and more consistent services. The report argues that the fastest path to scale is disciplined service design: bounded (clearly scoped, limited-risk)use cases, clear KPIs, and human oversight. It also highlights the risk of the “pilot trap,” where experimentation expands but production impact remains limited.
  • Procurement governance and contracting policy: The report shows that AI governance is also a contracting issue. Governments should use procurement to embed auditability, data portability, performance benchmarks, and remedies into AI systems from the outset. Trust must be written into contracts and vendor obligations.
  • Operational accountability, monitoring, and incident pathways: AI governance failures often emerge after deployment. The report calls for a shift to continuous assurance, including monitoring, named accountable owners, incident escalation, and redress mechanisms. In government, accountability must operate as a practical capability.

The report concludes with a roadmap for DCO Member States, setting out practical steps for governments at different levels of institutional maturity to build service capability, procurement discipline, and operational accountability as AI becomes a more embedded layer of the state.