In a move that underscored the complexity of municipal governance, the San Francisco Board of Supervisors decided to postpone voting on a sweeping ordinance aimed at overhauling the city’s code. The 362-page proposal, which sought to make well over a hundred changes, was put on hold after several supervisors and even the typically reserved board clerk expressed significant concerns.

The ordinance, proposed by Supervisor Rafael Mandelman and crafted with input from the city attorney’s office, was designed to streamline a city code that spans over 16 million words. The goal was to remove around 150 redundant sections and modify a few dozen others. However, the sheer scope of the changes raised eyebrows and prompted a closer look.

Concerns Over Clarity and Public Understanding

During a committee meeting on June 18, board clerk Angela Calvillo highlighted a critical issue: the legislative digest summarizing the proposed changes did not cover every alteration in the ordinance. This lack of transparency made it challenging for her and her staff to fully grasp the implications of the changes.

Calvillo emphasized that if someone of average intelligence found it difficult to map the changes onto the legislative text, perhaps part of the public would have a problem doing that as well. This concern was echoed by Supervisor Myrna Melgar, who admitted to turning to artificial intelligence to help her understand the sweeping changes.

I went home and I ran it through Claude Melgar confessed during the Tuesday board meeting. While she agreed with the spirit of the law, she identified specific changes that warranted closer scrutiny. For instance, she questioned whether the city still needed to track gender pay equity, highlighting the need for careful consideration of each proposed change.

The Role of AI in Crafting the Ordinance

The ordinance itself is a product of AI research, resulting from a collaboration between the city attorney’s office and the Regulation, Evaluation, and Governance Lab (RegLab) at Stanford University. The research group used AI tools to identify and analyze redundancies, which were then manually reviewed by city attorney staff.

Generally, the proposed changes aimed to eliminate or reduce reporting requirements which mandate that departments produce specific reports or information. Many of these requirements are outdated or obsolete. For example, the city code still requires the director of public works to produce a report every other year about pedestal zones for newspaper racks, which have long been removed. Similarly, the code refers to library fines that no longer exist since the library eliminated late fees.

Other reports are required at a frequency that the city attorney’s office deems excessive. Many of those reports end up being read by no one noted Board President Mandelman. Aside from the administrative burden, failing to produce these reports, no matter how unnecessary, technically violates the city’s code.

The Path Forward

In response to the concerns raised, Mandelman proposed postponing the vote to July 14 to allow for a thorough review of the issues. Melgar welcomed the delay, emphasizing the importance of transparency and public understanding. I think that as the legislative branch, we have a responsibility to our public to make sure that they understand what we’re doing. And I’m not sure that’s the case yet.

The postponement highlights the delicate balance between streamlining municipal code and ensuring that the public and legislators fully understand the changes being proposed. As San Francisco navigates this complex process, the role of AI in both creating and interpreting legislation will undoubtedly continue to evolve.