San Francisco is using artificial intelligence to fight against slow and inefficient government processes and excessive reporting rules. Supervisor Aaron Peskin is leading this effort. The city hopes to make its operations much smoother by removing or changing many of its current reporting rules. This big project aims to completely remove about 140 reporting requirements, which is more than a third of the nearly 500 requirements that can be changed through city laws.
The main reason for this effort is the understanding that too much paperwork takes away valuable time and resources, making it harder for the city to focus on important services for its residents. The huge amount of reports being created is seen as a major obstacle to good government, similar to problems seen at the national level. Peskin compared it to the millions of pages of paperwork Congress produces each year, calling it a “black hole” that pulls attention away from critical work.
With tight budgets, simplifying these processes is essential so city employees can focus on urgent issues. The solution to this problem comes from an AI tool created by a team at Stanford University, led by Professor Daniel Ho. According to Politico, this tool is built to read and understand legal text, acting like a highly advanced legal interpreter. The AI can also find outdated rules, which it does very efficiently.
San Fransisco is using AI to remove outdated policies
One example given was a rule requiring the Public Works Department to submit a report every two years about the city’s fixed newspaper racks. This rule is useless because those racks no longer exist. This is just one of hundreds of similarly outdated requirements that weigh down various city departments, including the controller’s office, the city administrator’s office, the planning department, and the Mayor’s Office of Housing and Community Development.
These departments were especially noted as struggling under what Ho called “policy sludge.” Before being used in San Francisco, the AI tool was tested and adjusted carefully. The team used it to find over 1,400 known reports in the U.S. legal code, along with hundreds more that had never been officially recorded. This thorough testing makes sure the AI’s analysis is accurate and reliable.
BREAKING: San Francisco Turns to AI to Clean Up Its Own Mess. Quietly Echoes Trump’s Deregulatory Vision
— Washington Eye (@washington_EY) June 5, 2025
In a move that sounds more like a confession than a reform, San Francisco has partnered with Stanford to use artificial intelligence to delete hundreds of outdated laws and… pic.twitter.com/xpndjIpESP
This isn’t the first major use of this AI technology. Last year, the same Stanford lab successfully used a similar AI tool to uncover racial restrictions hidden in millions of property records. The proposed law, which will be introduced to the San Francisco Board of Supervisors, already has support beyond Peskin.
Supervisor Bilal Mahmood has voiced strong approval for the plan, acknowledging how much unnecessary legal clutter burdens City Hall. The deployment of AI in governmental processes echoes themes explored in popular culture, like in A.I. Artificial Intelligence, Steven Spielberg’s most under-appreciated and misunderstood film, which delves into the complexities of artificial intelligence in society.
Mahmood said this law will greatly reduce bureaucratic waste, letting city departments focus on serving San Francisco residents directly. This bureaucratic overload isn’t unique to San Francisco; similar inefficiencies have led to unexpected situations elsewhere, such as a recent incident where a woman experienced a bureaucratic meltdown that led to her flashing fiasco.
