In a city known for innovation, San Francisco is bringing artificial intelligence into an unexpected arena: municipal governance. The city recently showcased how AI tools can modernize outdated legal frameworks that have accumulated over decades—a breakthrough approach that could revolutionize how local governments function nationwide.
Last month, I attended a fascinating demonstration at San Francisco City Hall where officials revealed how they’ve deployed machine learning algorithms to analyze and streamline the city’s municipal code. The initiative, part of Mayor London Breed’s broader government efficiency program, aims to identify redundancies, contradictions, and outdated regulations that hamper both business operations and resident services.
“We’ve had legal code accumulating since the Gold Rush era,” explained Jennifer Johnston, Deputy City Attorney, during the presentation. “Some regulations conflict with each other, others reference departments that no longer exist, and many simply don’t make sense in today’s world.”
The problem isn’t unique to San Francisco. Cities across America struggle with legal bloat—municipal codes that have expanded haphazardly for decades, creating compliance nightmares for businesses and confusion for residents. Traditional methods of code review involve armies of attorneys manually poring over thousands of pages, a process both expensive and painfully slow.
San Francisco’s approach leverages natural language processing (NLP) to scan its entire 1,600-page municipal code, identifying patterns and inconsistencies that human reviewers might miss. The AI system, developed through a partnership between the city’s Digital Services team and Stanford’s RegLab, can process the entire code in hours rather than months.
The system flagged over 400 instances of outdated terminology, nearly 200 contradictory regulations, and identified approximately 150 permit requirements that could be simplified or eliminated entirely. Perhaps most impressively, it mapped complex regulatory pathways that businesses must navigate, revealing bottlenecks where multiple department approvals create unnecessary delays.
According to research from the Brookings Institution, regulatory complexity costs American businesses an estimated $1.9 trillion annually in compliance expenses. For small businesses particularly, navigating municipal regulations represents a significant barrier to entry and growth.
“What makes this approach powerful isn’t just finding problems—it’s visualizing the entire regulatory ecosystem so policymakers can make informed decisions,” noted Dr. Daniel Ho, faculty director at Stanford RegLab, who helped develop the system.
The AI system isn’t making autonomous decisions about which regulations to keep or discard. Instead, it presents analysis to human officials who maintain final decision-making authority. This human-in-the-loop approach addresses common concerns about algorithmic governance while still leveraging AI’s processing power.
The initial results are promising. The city has already eliminated 32 redundant permit requirements and simplified application processes for small businesses. Restaurant owners, for example, previously needed to submit similar information to five different departments. Now, a streamlined process shares data across agencies, reducing paperwork by approximately 60%.
Beyond efficiency gains, the technology has revealed surprising insights about regulatory equity. The system identified areas where enforcement has been historically imbalanced across neighborhoods, potentially reflecting biases in how regulations are applied. This transparency creates opportunities to ensure rules are enforced fairly citywide.
“We’re not just cutting red tape—we’re making governance more transparent and equitable,” said Carla Short, Director of the Office of Digital Innovation. “When regulations are clear and accessible, everyone benefits.”
The approach isn’t without challenges. Privacy advocates from the Electronic Frontier Foundation have raised concerns about potential biases in how AI systems interpret legal language. Others worry about transparency in the decision-making process when complex algorithms are involved.
City officials counter that the system actually increases transparency by making the regulatory environment more comprehensible to both government workers and citizens. They’ve also committed to publishing methodology reports and maintaining human oversight throughout the process.
The technology also faces practical limitations. Certain regulations, particularly those involving complex value judgments or community standards, require nuanced human interpretation that AI systems cannot easily replicate.
Despite these challenges, other municipalities are taking notice. Representatives from Oakland, San Jose, and Los Angeles attended the demonstration, suggesting the approach may soon spread. The federal government is also showing interest, with the Office of Management and Budget sending observers to learn how similar technologies might streamline federal regulations.
If successful at scale, this approach could fundamentally change how cities govern. Rather than static, accretional legal codes that grow more complex over time, municipal regulations could become dynamic systems that adapt to changing circumstances while maintaining consistency and fairness.
For San Francisco residents and businesses, the benefits may soon be tangible: faster permit approvals, clearer rules, and more responsive government services. For other cities watching closely, it represents a promising path toward modernizing governance for the digital age.
As I left City Hall after the demonstration, I couldn’t help but appreciate the irony—artificial intelligence, often viewed as a disruptive force, might just help make our analog government systems work better for everyone. In a city where technology and governance haven’t always been comfortable partners, that would be an innovation worth celebrating.
