Business group calls for better coordination as California adopts AI

If California’s local governments, and the state as a whole, are to successfully adopt artificial intelligence, they need to make sure it is done in a more coordinated manner, according to a report released by a local business group.

The Silicon Valley Leadership Group, the leading business association for the region’s innovation economy, said in the report that, while many agencies at the state and local level are experimenting with AI tools, they are doing so without the capacity, procurement systems, data infrastructure or governance frameworks they need to properly evaluate those tools and ensure the technology’s use is valuable for their residents.

The group called on local governments to evaluate AI based on the problem it is addressing, the data it is using, the product or model being procured and the way it changes interactions between residents, staff and government processes. That evaluation goes far beyond asking whether an AI tool simply works, the report says.

“AI adoption in local government is not hypothetical; it is already happening,” said Jake Brymner, the author of the assessment and a policy fellow at the Institute for California AI Policy, which is housed at SVLG. “The question is whether agencies have the information, staff capacity, procurement tools and governance structures they need to support adoption. Procurement of AI tools is not just a technical exercise. It is a policy and implementation decision that can affect service delivery, worker roles and public accountability.”

California’s state and local agencies have been at the vanguard of AI adoption compared to many of their counterparts, and they have used the technology in a variety of areas. The state has embraced AI for its Covered California health insurance marketplace, while its judges are in the midst of testing an AI clerk for some cases. Meanwhile, its cities have turned to AI to speed up the permitting process to accelerate home construction and improve public transportation.

The state recognized there is more work to be done, too, and unveiled an effort to invite residents to help shape AI policy through Engaged California, its digital community engagement platform.

SVLG highlighted various good work already underway, including Los Angeles County’s use of AI to search through and summarize complex documents in public benefits programs. But the group also offered cautionary tales, like Long Beach’s chatbot pilot that reportedly returned outdated or inaccurate information to residents.

And if California is to truly embrace AI and use it to improve residents’ lives, the SVLG report said major challenges remain around procurement and deployment, especially at the local level. The group said agencies’ procurement systems “have not caught up with AI,” as they often receive inconsistent or incomplete information from vendors about model performance, training data, treatment of personally identifiable information, data use and long-term oversight.

Then, deployment needs “significant” change management, SVLG said. Many agencies do not have enough AI training, while they also grapple with uneven data governance practices, staff anxiety about automation and  collective bargaining considerations when AI changes workflows.

The report has a series of recommendations for local agencies, including that they designate AI policy leads to coordinate their efforts; create internal AI training programs; audit current AI uses and the data they use; publish internal guidelines and AI use inventories; update procurement methods to account for AI risks and disclosure needs; and share their ideas, best practices and lessons learned with their peers.

Meanwhile, the report urged the California Department of Technology to issue guidelines on information public agencies should receive when they procure AI tools. And more broadly, SVLG said the state should fund training for local agencies; create a state-supported sandbox to allow local agencies to test AI with synthetic data; pilot state data storage and compute services for local agencies; update data protection laws; and study how infrastructure can be scaled and financed.

“California has an opportunity to define what thoughtful public-sector AI adoption and deployment looks like,” Ziyang Fan, founding executive director of the Institute for California AI Policy at the Silicon Valley Leadership Group, said in a statement. “That means moving beyond fragmented experimentation and giving local agencies practical tools: shared standards, training, procurement guidance, testing environments and secure data infrastructure so AI can enhance public services while reducing risks.”

Despite the challenges that lay ahead, leaders said they are hopeful the state can continue to be a pioneer in AI adoption, and take advantage of that early movement to stay ahead of the curve.

“California is home to the world’s leading AI companies, and we are uniquely positioned to use these tools to expand human potential and solve real problems for Californians,” Ahmad Thomas, SVLG’s CEO, said in a statement. “Local governments are on the front lines of public service delivery. This assessment gives California a roadmap to help cities and counties adopt AI in ways that improve services, protect residents and build public trust.”