San Francisco, widely regarded as the global center of artificial intelligence, has been steadily introducing AI tools into city government to improve productivity. Since July 2025, around 30,000 city employees have been given access to Microsoft Copilot, an AI assistant powered by OpenAI. The technology is being used to help draft reports, summarize documents, analyze data, and reduce time spent on routine administrative tasks.
Despite the citywide rollout, AI adoption has been far from uniform. Departments focused on legal, analytical, and administrative work have embraced the technology the most, while employees in physically demanding roles such as police officers and firefighters make comparatively little use of it. The Department of Police Accountability has used AI to assist in reviewing body-camera footage, while the Department of Public Health has become one of the most active users for preparing communications and presentations.
Data from the city shows that by April 2026, employees had exchanged more than one million messages with Copilot. Although fewer than half of eligible workers in most departments actively use the tool, overall AI interactions have tripled since the program began. A relatively small group of "superusers" accounts for much of this growth, suggesting that enthusiasm and expertise with AI remain concentrated among certain employees.
To ensure responsible use, San Francisco has updated its AI guidelines by requiring human review of AI-generated content before it is used in public-facing or sensitive work. The city also prohibits AI from creating official documents without oversight or generating false or misleading information. Officials view the initiative as an evolving effort, with continued training and policy improvements aimed at expanding AI adoption while maintaining accountability and public trust.
