Artificial intelligence tools are gaining traction in the municipal industry and also raising regulatory and other questions, including whether big players' ability to deploy capital to build out AI tools could alter the muni market's structure by disadvantaging smaller rivals.
If building out AI tools requires a big investment, "it sure helps to have the capital," said Justin Marlowe, a research professor in the University of Chicago Harris School of Public Policy, adding that well-capitalized players such as bulge bracket banks, the biggest municipal advisors and large law firms will have "a built-in advantage" when it comes to AI.
"I think that's one thing the industry is going to have to contend with," said Marlowe, who also serves as director of the school's Center for Municipal Finance. Traditionally, the municipal industry has valued making sure that opportunities exist for smaller firms to have access to participate in deals, he said.
"And if access means having an LLM, … that may not be available to a lot of smaller shops," Marlowe said, referring to a type of AI system known as a large language model. "So, in addition to the kind of compliance considerations that the [Municipal Securities Rulemaking Board] and others are thinking about, you know there is that question of the implications for the structure of the industry."
Indeed, Marlowe isn't alone in thinking about AI's
"I think the adoption of AI raises some important questions from a regulatory perspective," Kim said during a panel discussion, adding that the MSRB's rules "are designed to protect issuers and investors in our market."
Among the MSRB rules he mentioned was Rule G-17. The rule "establishes a duty of fair dealing for your public finance bankers that serve you," Kim said, in a reference to issuers.
"And I think there's an interesting question about the structuring of your deals," he said. "If they're being done by an AI tool, are your bankers meeting their duty of fair dealing? Are they doing their due diligence to make sure that that structure is appropriate and meets your financing objectives?"
Another rule the MSRB CEO cited was MSRB Rule G-42, which pertains to municipal advisors.
"This rule establishes a fiduciary duty of MAs to their issuer clients," Kim said. "And if AI chatbots are actually the ones that are developing advice and making the recommendations that MAs are providing to their issuer clients, are MAs meeting their duty of care and upholding their fiduciary duty to their clients?"
Kim acknowledged that he didn't have answers to the AI-related questions he posed during the panel discussion.
"I think we're still trying to figure it out as an industry, and we're still trying to understand what the potential implications of AI are going to be … from a regulatory perspective," the MSRB CEO said.
For Marlowe, while he has ongoing conversations with muni industry professionals, his "main source of data" regarding the industry's use of AI comes from his students, who are studying for their master's degrees with an emphasis on municipal finance.
Information shared by students returning for the program's second year following summer internships or fellowships with industry firms has given Marlowe a window into how fast the muni industry's deployment of AI is moving.
For example, it was after students' summer stints in 2023 – the summer after OpenAI introduced the first version of ChatGPT in November 2022 – that he first heard students, particularly those who had worked at buyside firms, using terms like LLM or agent, Marlowe said.
"Summer of 2024, we had students for the first time reporting that they were … being given access to these tools and asked to experiment with them," he said, adding that students returning after having had jobs in the summer of 2025 reported consistently interacting with LLMs as part of their day-to-day work.
"So it's ramped up very quickly," Marlowe said.
Currently, AI is "being used across the industry, but I would definitely call its use uneven," he said.
"I think the most ubiquitous applications at this point on are the buyside," Marlowe said, adding that, while he declined to name them, he knows of "at least three or four big buyside firms" that have developed their own LLMs able to perform some of the day–to-day work traditionally done by buyside analysts.
"Obviously, there has to be human oversight, but a lot of that [work] can and has been automated," he said. "And it's allowed analysts to be able to … cover a larger number of credits, put more investment options in front of portfolio managers and, in some cases, to identify investment opportunities that might have otherwise been overlooked."
While the use of such AI tools can add value for the buyside, there's "a flip side," Marlowe said, adding that he's heard "rumblings of layoffs" and of firms hiring "fewer and fewer entry level folks" for roles where much of the work can be done by AI.
Consequently, a conversation he's been having a lot with students is "What does it mean to have a future in this industry if you don't have an opportunity to develop a lot of those kind of entry level skills by doing the work," Marlowe said.
"So, one of the things that we are contending with at the university level is whether we can provide more of those entry-level skill-building opportunities for our students before they go into the industry, knowing that they may have fewer opportunities to develop those types of skills on the job," he said.
Member firms represented by the Bond Dealers of America are looking at AI "from a lot of different perspectives," according to Michael Decker, senior vice president for research and public policy at BDA.
"And some of our biggest member firms are making big investments in it," Decker said, adding that some of BDA's medium-sized members "are kind of carefully watching how the AI landscape evolves before they dive in."
Among BDA members, "there seem to be a few areas where AI is emerging as a potential productivity tool," Decker said, citing document production and review as among those areas.
"And the other area where there's been a lot of activity is … on the trading desk – pricing," he said, adding that AI is being used "as an aid to traders in determining pricing."
Shayne Kavanagh, the senior manager of research for the Government Finance Officers Association, said that while GFOA hasn't conducted any surveys of its members to gauge how they might be using AI, "my observation is that most people are in the exploration phase of AI," meaning that "they're starting to try tools out" to see what they can do.
While Kavanagh stressed that he was cautious about making sweeping generalizations regarding AI use among GFOA members, a theme he pointed to was using AI to "solve the blank page problem" when initiating tasks.
"Generally speaking, that is where I'd observe most people to be at," he said.
