Berkeley City Council passes guidelines for future regulations on AI usage | City Government

The Berkeley City Council passed two resolutions that provide a framework for AI usage by the city during its regular meeting on March 10.

The first item, “The Berkeley Rule,” is a set of 10 guidelines authored by District 3 Councilmember Ben Bartlett that lay out the purposes that AI should fulfill within the city. The second item, authored by District 5 Councilmember Shoshana O’Keefe, included principles that the council would like the city’s AI policy to follow.

“The need to (create ‘The Berkeley Rule’) came from conversations with the city manager where he asked me to help come up with something to help him think about how to use AI, because it’s being used in a scattershot fashion,” Bartlett said. “He’s going to create administrative regulations right around it.”

The 10 guidelines that “The Berkeley Rule” outlines are: put residents first, modernize city services, empower the community, ensure transparency and accountability, standardize operations, certify ethical use, protect and prepare our workforce, defend civil liberties, social advancement and accessibility, and catalyze civic wealth.

The framework, which Bartlett has been working on for about a year, will guide how staff will regulate and incorporate AI systems into the city government.

“The administrative regulations are going to tell the city how (to use AI), but ‘The Berkeley Rule’ tells us why,” Bartlett said. “This has got to be a human-centered thing that really helps us fulfill the moral evolution of our city.”

While there has not been a formal policy that dictates how city staff should use AI, the second item is intended to inform the city manager of seven key principles to take into consideration when developing the policy.

The item complements “The Berkeley Rule,” according to District 4 Councilmember Igor Tregub, who co-authored “The Berkeley Rule” and co-sponsored the second item.

Policies include creating safeguards against AI systems introducing bias, protecting data privacy and ensuring cybersecurity compliance, as well as maintaining human oversight and accountability. In addition, the resolution also seeks to explore opportunities to integrate AI in operations management and foster cross-departmental collaboration of AI knowledge.

“Both items were crafted in close coordination with the city manager, and city staff is in the process of developing its own guidelines for the use of AI,” Tregub said. “(The item) directs (the city manager) in the writing of such a policy to keep certain principles in mind, which are the seven principles that you see in the item.”