Local agencies move to set rules as AI use expands in public sector

Belvedere employees must protect confidential information, disclose when they use AI and rely on human judgment under a new city policy.

The council unanimously adopted the policy as part of its consent calendar May 11, establishing formal rules for how artificial intelligence tools may be used in city work.

According to a staff report from City Manager Robert Zadnik, it was created to address the growing use of AI in workplace software and to prevent AI use “without clear standards, particularly with respect to confidential information, public records, accuracy, transparency and public trust.” City governments around the world have begun using AI, including large language models like ChatGPT, in recent years. A 2024 ThoughtLab study conducted in partnership with Deloitte found that 56% of the 250 cities surveyed were actively using AI, either selectively or widely, a share that was projected to reach 83% by 2027.

The Transportation Authority of Marin adopted an AI-use policy last month, basing it on guidelines from the Government AI Coalition, a network of public agencies led by San Jose. At its April 20 meeting, the Belvedere-Tiburon Library board approved a draft AI-use policy, pending legal review. The county of Marin adopted AI-use guidelines in May 2025.

Meanwhile, Tiburon remains without an AI-use policy.

Read the complete story in our e-edition, or SUBSCRIBE NOW for home delivery and access to the digital replica.

Comment on this article on Nextdoor.