The US federal government has disclosed 3,611 active or planned AI use cases across agencies, a 70% increase from the previous administration, covering sensitive areas including public health, individual freedoms, and nuclear safety. This represents a significant and largely unscrutinised transfer of consequential decision-making from humans to automated systems. The scale and pace of deployment raises serious concerns around accountability, auditability, and the absence of robust AI governance frameworks.

Security Architect’s Take: Cloud security architects supporting US federal workloads or supplying AI services to government should urgently review their AI governance posture — ensure audit trails, human-in-the-loop controls, and explainability requirements are baked into solution designs, particularly for high-impact decision workflows. UK and EU architects should also monitor this as a precedent that may influence domestic public sector AI adoption patterns.

Original advisory: AI Use by the US Government