Pierce Freeman and Richard Diehl Martinez speak with Harrison MacRae, Pennsylvania's Director of Emerging Technologies. Speaking in a personal capacity rather than for the state, he argues that AI policy should begin with the problem being solved, the people affected and the public mission.
The conversation contrasts federal and state approaches to regulation, including attempts to restrict state AI laws and targeted protections concerning children, discrimination, privacy and likeness. The participants debate the benefits and costs of a regulatory patchwork; their discussion is a policy snapshot from February 2026, not a comprehensive statement of current law.
MacRae explains why procurement can be slower than private-sector product development. Public bodies must consider statutory obligations, service continuity, vendor stability and accountability alongside experimentation. The hosts ask whether smaller pilots and more flexible approaches could make adoption more responsive.
Turning to implementation, they discuss build-versus-buy decisions, staffing constraints and the importance of reliable underlying data. They explore local deployment and agentic tools as possibilities, while stressing that data cleanup, interoperability and mission-specific safeguards remain central to useful public-service systems.
The closing discussion distinguishes long-term AI risks from immediate service-delivery needs and invites technologists to contribute to civic work. MacRae describes direct government roles, volunteer projects and fellowship routes, presenting public service as challenging but potentially meaningful work.
Watch on YouTube
