July 21, 2025
Shaping AI in State Government
SACRAMENTO, Calif. – On June 26, the California Department of Technology hosted its monthly Artificial Intelligence Community (AIC) meeting that brought together more than 200 state employees across 67 departments and agencies. The AIC is led by the State’s Chief Innovation Technology Officer, Vera Zakem. The June event featured guest speaker Dr. Radha Plumb, Former Chief Digital and AI Officer at the United States Department of Defense, who shared her perspective on scaling AI in practice.
Dr. Plumb has been working as a Senior Fellow at the Wharton School at University of Pennsylvania for the last few months in their accountable AI lab. She has been analyzing AI adoption in the context of what she learned at the Pentagon and how that can translate to large enterprises. To the surprise of the group, Dr. Plumb referenced a recent McKinsey study that indicates over three-fourths of organizations have figured out how to adopt AI in only one business function and just one-fourth of those are seeing bottom line impact. Put another way, this emerging technology is still very much emerging, and there is a lot to be learned before its adoption.
She outlined a shift from thinking about AI as a technical capability to thinking about how to conceptualize AI’s adoption – what should the technology be asked to do? From the former lens, organizations are confronted with a false choice between human judgment and machine capability: either AI is adopted, and the worry is centered around if the AI is acting ethically, or AI is not adopted, and the performance gains are lost.
At the Pentagon, she abandoned this human-versus-machine mindset in favor of thinking in terms of a human-machine team. Then came the next challenge of integrating AI systems into the human process so that the human-machine collaboration could exceed the performance of either working alone. Dr. Plumb distilled this challenge to three foundational considerations to address:
- Infrastructure: Ensuring access to the necessary computational resources, cloud architecture, data and integrations to effectively leverage AI.
- Governance: Understanding how a tool is used, its performance, bias and fairness, compliance with legal liability risks, cyber compliance and organizational responsibilities.
- Workforce: Ensuring the workforce is effectively prepared to integrate and collaborate with machines as a human-machine team.
Dr. Plumb emphasized a key takeaway for the AIC was to think about infrastructure, workforce, and governance process at the beginning of AI initiatives to effectively iterate during the build and test phases and eventually integrate and scale for successful implementation. Her experience of leveraging 90-day sprints where technology users worked alongside technology builders, enabled delivery of a minimum viable capability in about 18 months that was then scaled globally just six months later. She encouraged baking in these core pillars of infrastructure, workforce, and governance process at the beginning of the journey in order to “train to them, build to them, invest in them.”
To close out her presentation, she noted, “California… has tremendous potential to lead the way on how AI adoption should look, especially in the public sector.”
Following the presentation, CTIO Zakem engaged Dr. Plumb in a series of questions that delved deep into technical subjects, exploring methods for measuring impact and governance. The conversation broadened to more reflective themes, where Dr. Plumb shared what she was most proud of and discussed the areas she wished she had more time to address during her time at the White House.
The meeting concluded with a lively Q&A session for all participants. These inquiries spanned a range of critical topics for the state, including procurement barriers, how to manage AI risks, and the application of AI in cybersecurity.
The AI Community is open to California State Employees. To join the community, state employees can subscribe to the AIC listserv.