California's New AI Sentinels: Embedding Experts to Govern the Future

SACRAMENTO, CA – July 07, 2026

In a move that signals a profound shift in how governments intend to manage the complexities of artificial intelligence, the California Council on Science and Technology (CCST) today announced the launch of its AI Science Residency Program. The initiative places top-tier AI experts directly within the state's key agencies, effectively embedding scientific sentinels on the front lines of policy and public safety. The two inaugural advisors, Michael Chen and Dr. Justin Norman, have begun their residencies at the Governor's Office of Emergency Services (Cal OES) and the Department of Technology (CDT), respectively, tasked with advising on the safety and governance of "frontier AI"—the most powerful and unpredictable systems being developed today.

This is more than a simple staffing announcement; it's a strategic masterstroke. As the global epicenter of AI development, California is not just reacting to the technology's rapid evolution but is actively building a new apparatus for governance. By embedding independent expertise directly into the machinery of the state, California is creating a model that could define how democracies grapple with technologies that are outpacing traditional oversight.

From Legislation to Live Operations

The timing of this initiative is no coincidence. It arrives just months after the state's landmark Transparency in Frontier Artificial Intelligence Act (SB 53) went into effect on January 1. This pioneering law—the first of its kind in the United States—mandates that developers of the largest AI models conduct catastrophic risk assessments and, crucially, report any critical safety incidents to Cal OES. The law created a framework, but it left a critical question unanswered: who within government possessed the requisite expertise to evaluate these highly technical reports and advise on a response?

The AI Science Residency Program is the answer. It provides the human infrastructure needed to translate legislative intent into operational reality. “As frontier AI development and governance rapidly evolve, good policy relies on a deep partnership between governance officials and cutting-edge technical expertise,” said Julianne McCall, PhD, CEO of CCST. “This program connects California agencies with the top-tier AI talent they need to evaluate and advise on emerging risks, helping ensure technology serves the public good.”

This partnership aims to bridge the dangerous chasm between the speed of technological innovation and the methodical pace of bureaucracy. By placing advisors directly within the agencies responsible for implementation, the state is ensuring that its response to an AI-driven crisis or a legislative query is informed by current, ground-truth scientific understanding, not by outdated briefs or industry-provided talking points.

The Advisors on the Front Line

The success of this initiative hinges on the caliber and focus of the advisors themselves. The inaugural placements demonstrate a remarkably strategic alignment of expertise with agency needs.

Michael Chen joins Cal OES, the agency now responsible for receiving and analyzing safety incident reports from frontier AI developers under SB 53. His background is a perfect match for this daunting task. Chen previously worked at METR, a nonprofit focused on evaluating autonomous AI agent capabilities and risks, where he advised leading AI developers on mitigating catastrophic threats and even reviewed a developer's report on sabotage risk from AI agents. His work at Cal OES will focus on exactly these issues: critical safety incidents, AI-powered cyber defense, and the risks posed by developers' own internal use of AI.

"AI agents are learning to autonomously execute exponentially more complex projects over time," Chen stated, highlighting the urgency of his mission. "That now includes cyberattacks and will pose risks for critical infrastructure." His role is not merely academic; it is to help California build proactive response plans for threats that are, for now, largely theoretical. As Commander Matthew Sage of the California Cybersecurity Integration Center (Cal-CSIC) noted, Chen's expertise “will meaningfully strengthen our capacity to respond to the rapidly evolving frontier AI risk landscape.”

Simultaneously, Dr. Justin Norman has been placed within the California Department of Technology (CDT). His role is to be the state's chief technical translator, advising leadership on how the latest AI developments should inform the definitions and thresholds within California's safety legislation. With over 20 years of leadership experience spanning the Defense Innovation Unit, Yelp, Cisco, and the U.S. Marine Corps, combined with a recent PhD from UC Berkeley focusing on the robustness of generative AI and deepfake detection, Dr. Norman is uniquely positioned to bridge the worlds of advanced research, commercial application, and public policy.

"California has the opportunity — and the responsibility — to demonstrate that governments can produce effective technically and socially grounded AI policy that still supports rapid innovation," Dr. Norman said. His task is to ensure that as AI models evolve, the laws designed to govern them do not become obsolete, providing the technical clarity needed to make legislation like SB 53 enforceable and effective.

A New Model for Governance

Perhaps the most innovative aspect of the program is its structure. The advisors are formally employed by CCST, a nonpartisan nonprofit, and placed in their host agencies via a memorandum of understanding. This model, inspired by the role of the Science Advisor to the President, is designed to ensure the advisors maintain an independent scientific perspective while being fully integrated into leadership teams. They are insiders with an outsider's objectivity.

This structure is a deliberate attempt to solve a core problem of modern governance: attracting and retaining elite technical talent that can command far higher salaries in the private sector. By creating a prestigious, mission-driven residency, CCST offers a compelling alternative, allowing top minds to contribute to public service without becoming enmeshed in the traditional civil service structure.

Government Operations Secretary Nick Maduros expressed his enthusiasm for the partnership, stating, “We’re looking forward to working with these incredible advisors to help the state navigate emerging issues around artificial intelligence.” This sentiment underscores the recognition at the highest levels that the state cannot go it alone.

California's experiment is a blueprint for a new kind of technocratic governance—one where scientific expertise is not just consulted but is woven into the very fabric of the state's operational and strategic decision-making. As the state government itself becomes a user of AI, with tools like its own custom generative AI, "Poppy," being deployed across departments, the need for this internal expertise becomes even more acute. The advisors are there not only to regulate the industry but to help the government regulate itself. As other states and the federal government watch closely, California's AI sentinels are not just protecting one state; they are testing a model for how to secure the future for everyone.