
Public Sector AI Is Shifting Toward Smaller, Tighter Models Built for Control
A new MIT Technology Review Insights report argues that government agencies may be better served by purpose-built small language models that can operate under tight security, governance, and connectivity constraints.
- The report argues that small language models may suit government better than cloud-heavy large-model deployments.
- Security, governance, limited connectivity, and infrastructure constraints are central barriers.


