A (4:08)

We've talked a lot to your point about the implementation of Gen AI. Like how should that happen? There are a lot of different paths. And you spoke about this with Carrie Bishop and Michelle Haynes actually in our very last episode. And then earlier in the season you talked about this with Oliver Wise at GovX @Johns Hopkins. If you were mayor again, how would you be implementing Gen AI, leading the implementation of it, training your employees on Gen AI use?

B (4:38)

Right now, I'd love to be mayor again. Now the tools are just. Even though the problems are greater than ever before, the tools are just incredibly better. They're even better than when I worked for Mike Bloomberg a decade ago or so. Let me start with answering your question this way. I don't think I would do generative AI training. What I would do is identify problems that if we applied generative AI to the data, would solve the problem, identify processes that could be changed through the use of generative AI. When in the early days of the data analytic effort across the country, Bloomberg Philanthropies funded several data analytics centers, if you will, one in Chicago with very talented person in charge. And that began as an initiative around data and it changed its approach to a initiative around problems that could be solved with data. So here I would say that broad based approaches of generative AI to use data to solve problems, it may be the deputy Commissioner of Transportation. It may be the assistant social worker for the north side of Chicago, working for the state or the county, but I would identify the problem and the user. Then I would look at how the data can be organized and clean so it's available. And then I would train that person on how to use generative AI for better and more success.

A (6:10)

So it'd have to be identifying the problem first and then knowing what could be done about that problem. With Gen AI as a tool, but not trying to force Gen AI in first. It sounds like yes.

B (6:23)

I mean, it's hard to train people who are busy all the time, every day on tools without them understanding how it's going to change and improve the way they do work. So the capacity for change is dramatic. It's huge. But I would start again with, okay, let's think about problems you face. Like, all right, you're transportation Assistant Commissioner and you're trying to figure out what 80, 20 rules apply. Why are 80% of your potholes the same place year over year? What's causing those conditions? How can you acquire the data using generative AI tools to help you understand how to solve that problem or do the same in social services? It's a combination of the two. I mean, I've kind of overstated it, but I would do it that way.

A (7:15)

Something else that we've been talking a lot about that is related to Gen AI, but is also a much wider spread issue is a topic of trust. We know that trust in public institutions has been steadily declining over the past several years. And you recently wrote about the real world impacts this is having on public servants. So what are some things that could help address this trust gap and maybe potentially start to close it? Whether that's policy, technology, data, all of the above.

B (7:49)

Well, these are fun questions because I actually think there's some answers to these questions. People are frustrated with their government more at the national level than the local level, but they're frustrated at the local level too. Trust is much higher at the local level, is at the national level, but it's still not going up and it's kind of subsiding a little bit. So how can you increase trust? Well, you can increase trust by having your residents understand that you're listening to them and you're doing best to solve or prevent or preempt the problems they face in their day to day lives that the mayor may have some control over. So there are digital tools to allow you to listen better. Right? You can do social media, you can do sentiment mining, you can do polling or you can actually do virtual conversations with a much broader array of community residents than just the traditional meeting. Then you can take those problems, prioritize those problems, and use data then to determine what's causing those problems and solve it. So just to make a complicated answer sound a little bit easier, it's a virtuous cycle. Right. You listen better, you listen more comprehensively, you react quicker, you react with a way that allows you to use the data to preempt and solve a problem. Instead of just responding in a late way, you communicate that to the resident, you build up little buckets of trust. So the opportunities, generative AI are much better in terms of responding. But eventually, and we have a couple projects like this at our center at Harvard, the Bloomberg center for Cities, Generative AI will also soon allow communities to understand their conditions better because they'll be able to access open data and understand open data in a way today that is a little bit burdensome.

A (9:38)

It's interesting because you're saying sort of a similar thing for the employees using genai, starting with a problem and addressing that and then also having the residents start with a problem that they have. And I think you always kind of say that the residents, they know the problem and they have a sense of the solution. It's just kind of like, how do you get that in? So I can see that being a path for Genai.

B (10:01)

Yeah. I mean, the city officials have lots of data and very little context. And the community has a lot of context and very little access to the data. And put them together, we should have a better solution.

A (10:11)

Yeah. All right, so we are entering the lightning round. So I'm going to ask you some rapid fire questions.

B (10:16)

Lightning round.

A (10:17)

I want the first thing that comes to mind. So what was something that you really enjoyed learning about this year?

B (10:25)

Oh, I think learning about generative AI. I mean, I'm a public official, local official, I'm a translator. Right. Between technologists and cities. And so you can read in the popular press about generative AI, but reading the technical articles. Right. Understanding kind of how it could change so dramatically. So I have a long way to go, but I think that subject matter has been fascinating to me.

A (10:53)

Yeah, I agreed. If you could summarize 2025 in one word or phrase, what would it be?

B (11:01)

Difficult challenges, too much anger. But great promise because of what we could do.

A (11:10)

You know, that wasn't a word, but it was a good phrase.

B (11:13)

I liked it.

A (11:13)

I liked it.

B (11:14)

First you said a word and then you said or a phrase. I heard you.

A (11:17)

Well, I Think it was extremely accurate. Do you have a book, a paper or an article or maybe a talk? Anything to recommend to the listeners from this last year?

B (11:30)

Well, I just came from a meeting with Jennifer Palka and we had her as a podcast. And so that just top of mind, I think everybody should read Recoding America why Government's Failing in a Digital Age. It's a good prescription about how we can change things for the better.

A (11:46)

Yeah, definitely. What is something that you're particularly excited about that we are working on at DataSmart at the center?

B (11:54)

Well, we've got several things that are really exciting. As you know, we work with most of the large cities in the country and the people who attend our meetings, the chiefs of staff and deputy mayors and even mayors are just such a committed group that working with them is inspiring. A couple of the projects that we have, one with the Knight foundation that looks at how generative AI can improve community engagement with cities in a way that helps solve problems, is essentially a project around your question, which is how does responsiveness produce trust? Very exciting. And another one which just shows how much we can do with data and with committed individuals would be where are there hyper local environmental conditions that create public health problems, say, for anybody, but with the concentration on children and about which the city can do something? So how can they identify the causes of chronic asthma or other issues associated with some environmental condition and do something about it? Because now we have the data. We have the ability to understand the data and the ability to take action. Licensing, permitting, inspecting. I think both of those are really quite interesting areas of inquiry.

A (13:06)

So you already mentioned Jennifer Palka, but who else do you think our listeners should also be following? Anyone that you found particularly inspiring this year?

B (13:17)

The people who are most inspiring are the ones I've talked to in the last, like two weeks, because I don't remember who they are. But, you know, I would follow Bill Eggers, who's in charge of research at Delaware Deloitte has a great comprehensive list of literature and practical examples in this area. Mitch Weiss at the Harvard Business School is one of the leading thinkers on the application of generative AI to cities. I came from a meeting at Cornell Tech. They're doing a lot of really interesting work in New York City with their digital fellows. And then nearby mit, Sensible City Lab is providing really good information. All of those are sources of ideas and, and practices.

A (13:54)

And I have to recommend, speaking of Mitch Weiss, that we've had him on the podcast before. And then we're also going to be bringing him back in the next year as a guest because he's doing a lot of really great work on this. What surprised you in 2025?

B (14:11)

Well, 2025 was an interesting year because there was so much dynamic change. And dynamic change can lead to a bunker mentality, or dynamic change can lead to. Well, we really need to innovate because we can't keep doing things the way we did before. And what surprised me a little bit is we haven't gotten to the. Well, these are the conditions for bold, innovative change. We're still kind of stuck in the responsiveness mode. So I think as we go into 2026, it'll be an opportunity to kind of rethink the processes of government, to make the best of the current conditions.

A (14:45)

That was our next question. What do you hope to see? So it seems like you hope to see people kind of seizing the reality that they're in right now and finding ways to kind of advance even with the barriers that are existing right now.

B (14:59)

Yes, these digital tools can't be cute things that we use on top of a legacy government. We need to change the legacy government dramatically to give employees more discretion, to give them more information, to give them the capacity to solve problems that they really want to solve if they had the flexibility and data to do it. We need to think broadly about the innovative structures of government. How can we change outdated ways we do things to allow much more effective delivery of services. How do we use the data to make government intervene where it can make the most difference? You know, like return on investment? Where should the dollars be spent? And we can't have this generative AI be like this cute little thing that we do at night to find out what movie to watch. It has to be incorporated in the processes of government, which means government has to spend more on training. It never does spend much on training. It needs to liberate its employees. It needs to give them the tools, and it needs to use the evidence to work better. If you do all of those things, I really think as difficult as 2026 might be for cities, it could be a watershed for change in the way cities operate.

A (16:07)

Well, thank you so much, Steve, for having me on from behind the screen so that I can interview you. And yeah, fingers crossed. We see that in 2026.

B (16:17)

Thanks for your time.

A (16:18)

Thank you. If you like this podcast, please Visit us at datasmartcities.org and find us on itunes, Spotify, or wherever you get your podcasts. This podcast was produced by me, Betsy Gardner. Thanks for listening.