Check out the episode:

You can find the shownotes through this link.

---

Are you interested in the evolution of AI use in urban governance?

---

Our summary today works with the article titled Artificial intelligence and the local government: A five-decade scientometric analysis on the evolution, state-of-the-art, and emerging trends from 2024, by Tan Yigitcanlar, Sajani Senadheera, Raveena Marasinghe, Simon Elias Bibri, Thomas Sanchez, Federico Cugurullo, and Renee Sieber, published in the Cities journal.

This is a great preparation to our next interview with Thomas Sanchez in episode 330 talking about urban and technological evolution.

Since we are investigating the future of cities, I thought it would be interesting to see how AI is utilised in urban governance. This article investigates the evolution, current state and emerging trends of AI in local governments across 5 decades.

[intro music]

---

Welcome to today’s What is The Future For Cities podcast and its Research episode; my name is Fanni, and today we will introduce a research by summarising it. The episode really is just a short summary of the original investigation, and, in case it is interesting enough, I would encourage everyone to check out the whole documentation. This conversation was produced and generated with Notebook LM as two hosts dissecting the whole research.

---

[music]

Speaker 1: we hear so much about artificial intelligence, AI changing everything, what’s actually happening closer to home. I. Like in our local governments.

Speaker 2: Right. It’s definitely moving beyond just ideas. AI is starting to actively shape how our cities work,

Speaker 1: how things like the services they offer and decisions they make.

Speaker 2: Exactly that. And to really understand this, we looked at, we have 7,000 research papers. 7,000, yeah. Spanning five decades too. From 1973, right up to October 20, 23. So it gives us this comprehensive sort of data backed view of the science behind AI in your local community.

Speaker 1: Okay. So this isn’t just stories we’ve heard. You used a method science to metrics, you said to see the big picture in all that research.

Speaker 2: Precisely. It lets us see the major patterns, the trends, what the whole research field is focusing on. It gives us some pretty surprising insights into where AI is making the biggest impact right now.

Speaker 1: Okay. Let’s dig into that. What’s the main takeaway when you look at all that data?

Speaker 2: The first thing that really jumps out is just how well. New. This all is in the grand scheme of things. How new? If you look at the timeline, the first 30 years we study, that’s 1973, all the way to 2004. Barely 1% of all the research on AI and local government was published then

Speaker 1: 1% over 30 years.

That’s. Tiny. It really drives home that this is a recent shift, like the starting gun only just fired.

Speaker 2: It really does. Something clearly changed around 2004.

Speaker 1: So what happened then? What was the catalyst

Speaker 2: after 2004? It just accelerates massively. Like 99, 9% of all the research we looked at came out since then.

Speaker 1: 99%.

Speaker 2: Yeah. If you think about the wider tech world, that’s when we saw the huge explosion of the internet, social media taking off the whole big data era starting.

Speaker 1: Ah, okay. So suddenly there was all this data and maybe the computing power to actually do something with it.

Speaker 2: Exactly. Those advancements created both the fuel, the huge amounts of information and the engine, the powerful computing tools and algorithms that AI needs to really work.

Speaker 1: So that tiny trickle turned into a flood. It wasn’t just a steady increase, was it? I remember you mentioning a more recent surge.

Speaker 2: That’s right. The growth has been particularly intense just in the last decade or so, say from 2014 up to now. That really highlights the current boom, the real push for practical AI applications at the city level that we’re seeing today.

Speaker 1: Okay, so AI is definitely gaining ground fast in local government, but why? What are they actually trying to achieve with it? What problems is it solving?

Speaker 2: The research really points to four main reasons for key purposes. There’s decision support, automation prediction. And improving service delivery for you, the citizen.

Speaker 1: Right. So helping officials make better choices, handling routine stuff automatically, maybe seeing what’s coming down the road and making services better.

Speaker 2: That’s a great summary. They’re using AI to get more insights for decisions, make routine tasks more efficient, anticipate future needs or problems, and ultimately, yeah, provide better, faster services.

Speaker 1: Makes sense. Can we get more specific, what are some concrete examples from the research for each of those? Let’s start with decision support. Sure.

Speaker 2: Some early examples. There were things called expert systems. These were designed to help with quite complex tasks like figuring out land use, planning options, or making sure building plans, met all the codes.

Speaker 1: So like an AI assistant providing data-driven advice,

Speaker 2: offering insights to help the human officials make the final call.

Speaker 1: Okay. And automation, that’s a word we hear a lot. How does that look in city hall?

Speaker 2: In this context, it often means using what some call algorithmic bureaucracy. Basically using algorithms for tasks like sifting through huge data sets or allocating resources based on defined rules.

Speaker 1: So it takes over some of the grunt work.

Speaker 2: Exactly. It frees up the human staff to focus on more complex problems, things that need human judgment or interaction.

Speaker 1: Got it. What about prediction? How are cities using AI to look into the future?

Speaker 2: We’re seeing more and more use of AI for forecasting things like predicting traffic jams so they can adjust traffic lights or forecasting energy demand to manage the grid better, and even trying to predict where public safety issues might arise, though that’s obviously a complex and sensitive area.

Speaker 1: Definitely. Okay, and the last one, improving service delivery for us.

Speaker 2: Yeah. A really common example now is chatbots. Lots of cities are using AI powered chatbots on their websites or apps.

Speaker 1: The little popup windows asking if you need help.

Speaker 2: Exactly. I. If they can give you quick answers to common questions about, say trash pickup schedules, or how to pay a bill, guiding you to the right info without you having to wait on hold or search through pages. Los Angeles has one called CHIP, for instance.

Speaker 1: Interesting. So AI is really touching a lot of different parts of how local government operates. If we zoom in a bit more, what are the specific sectors or areas where it’s being applied most?

Speaker 2: The research really highlights a few key zones. Planning is a big one. Analytics making sense of data security, surveillance, energy, and also modeling.

Speaker 1: Okay. Planning. We mentioned land use earlier. What else falls under planning?

Speaker 2: Things like optimizing public transport routes, figuring out the best locations for new facilities, or even simulating how a new development might impact traffic or local services

Speaker 1: and analytics. I guess cities must have mountains of data, right?

Speaker 2: Oh, absolutely. So machine learning and big data techniques are huge here. They used to spot trends in everything from public health issues and disease spread to understanding infrastructure, wear and tear, helping make those decisions more evidence-based.

Speaker 1: Security seems crucial. How is AI playing a role there?

Speaker 2: There is a definite growing focus in the research on cybersecurity using AI to detect hacking attempts or unusual network activity. Also, looking into things like attack detection more broadly, and even exploring potential uses of blockchain for data security.

Speaker 1: Okay. Cyber threats make sense. What about physical security or surveillance? That can be controversial.

Speaker 2: It can be, and the research reflects growing use, particularly of computer vision. That’s AI analyzing images or video

Speaker 1: for what kinds of things?

Speaker 2: Things like monitoring traffic flow, automatically identifying potential fire hazards from camera feeds, tracking pedestrian activity in public spaces, or even assessing the condition of roads or bridges from images.

Speaker 1: Right. Okay. Energy is another big one for cities. How’s AI helping there?

Speaker 2: It’s being used to optimize energy use in city buildings. For instance, adjusting heating and cooling automatically. Also, managing smart grids more effectively, balancing supply and demand, and even sometimes in programs to encourage residents to conserve energy.

Speaker 1: And the last area you mentioned was modeling. What does that involve?

Speaker 2: This often involves using ai, especially techniques like convolutional neural networks, CNNs, which are really good at finding patterns in complex data like images or sensor readings.

Speaker 1: So using them for what?

Speaker 2: For things like forecasting how air pollution might spread, modeling the potential impact of climate change on city infrastructure, or predicting how changes to the road network might affect congestion.

Speaker 1: It’s quite a range. Now, you mentioned this research covers decades. Did you see distinct phases or eras in how AI was approached by local governments over time?

Speaker 2: Yes. The analysis suggested roughly four phases. There was a kind of foundational period, maybe late eighties to mid nineties. That’s when those first expert systems started appearing mainly in planning.

Speaker 1: Okay. The very beginning.

Speaker 2: Then a conceptual phase, say mid nineties to the mid two thousands, that’s when the term artificial intelligence itself started showing up more prominently in the research, but maybe less practical application yet.

Speaker 1: More theory and discussion a Speaker 2: bit, yeah. Followed by an exploratory phase, mid two thousands to mid 2010s. Here you start seeing the first real clusters of research around AI for specific things, planning again, but also analytics, energy, mobility starting to emerge.

Speaker 1: Okay, starting to get more focused.

Speaker 2: And then finally, the phase we’re in now, the booming phase from the mid 2010s onwards. This is where the idea of the smart city really takes hold. Machine learning becomes the dominant approach.

Speaker 1: Right. That connects back to the massive growth you mentioned earlier. Yeah. So the focus shifted from general AI concepts to very specific tools within this smart city idea.

Speaker 2: Exactly. Much more application focused now.

Speaker 1: So we’ve looked back and looked at the present. What about the future? What emerging trends does the research point to? What’s next for AI in our cities,

Speaker 2: the research shows growing momentum or heat. Around a few key areas, security, mobility, surveillance, and modeling are definitely getting more attention.

Speaker 1: Okay. Can you give some examples what’s happening in security?

Speaker 2: There’s definitely more research looking at cyber threats as we mentioned, but also exploring things like blockchain for secure data management in government services

Speaker 1: and mobility. That seems like a big one for cities.

Speaker 2: Huge. We’re seeing lots of innovation here. Think intelligent transportation systems that adapt in real time. More research related to autonomous vehicles and how cities manage them. Smart parking solutions, even creating real-time digital twins of the city’s transport network to test changes.

Speaker 1: Digital twins. Like a virtual copy

Speaker 2: essentially? Yeah, a simulation tool. And in surveillance, more advanced computer vision applications, detecting specific events. Differentiating between pedestrians and vehicles more accurately, maybe using drones for monitoring certain situations and modeling. You

Speaker 1: mentioned CNN’s earlier.

Speaker 2: Great. The use of those deep learning models like CNN’s is growing for all sorts of urban challenges, and there’s also this interesting new-ish field called geo AI popping up geo ai. Yeah. Basically applying AI specifically to geographic data, spatial data for. Better. Urban analysis, environmental monitoring, planning, that sort of thing.

Speaker 1: It also feels like we should mention things like chat, GPT, the large language models. Are they showing up in the research for local government?

Speaker 2: That’s a great point. Interestingly, while generative AI like chat, GPT obviously have huge potential to say, revolutionize how citizens interact with government information or services, it’s still very new in the academic research focus specifically on local government applications.

Speaker 1: So the research hasn’t quite caught up with the hype yet,

Speaker 2: not fully, at least in the papers published up to late 2023. You can absolutely see the potential there. For things like more natural chatbots summarizing complex regulations or even helping draft communications, it’s likely an area we’ll see explode in the research very soon.

Speaker 1: Something to watch. I. Now with all this focus on technology and applications, is there anything the research suggests is being overlooked?

Speaker 2: Yes, definitely. And this is really important. There’s a significant gap when it comes to research on the ethics, the public engagement aspects, and the accountability surrounding AI use by local governments.

Speaker 1: So lots of focus on what AI can do, but less on how it should be done responsibly.

Speaker 2: Precisely terms like AI ethics or responsible AI have started appearing, but it’s very recent and still a relatively small part of the overall research landscape we analyze.

Speaker 1: It seems like a pretty big blind spot, doesn’t it? Especially given the potential impact on people’s lives.

Speaker 2: It really does. There’s some work on what’s called explainable AI or X ai, trying to make AI decisions less of a black box

Speaker 1: so we can understand why the AI made a certain prediction or recommendation.

Speaker 2: Exactly. It’s a step towards transparency, but the practical application and effectiveness of X AI in complex real world local government settings that still needs a lot more investigation according to the literature.

Speaker 1: A question maybe for you, the listener to think about is, yeah, as AI gets woven deeper into the fabric of our cities, how do we ensure it actually benefits everyone and doesn’t create new problems or biases?

Speaker 2: That’s the critical question, really.

Speaker 1: Okay. If we boil down this huge analysis, what are the absolute key takeaways?

Speaker 2: I’d say number one. AI and local government isn’t science fiction anymore. It’s here and it’s growing exponentially fast.

Speaker 1: Decades of slow burn then boom.

Speaker 2: Exactly. Number two, its main uses right now are really focused on helping make decisions, automating processes, predicting outcomes, and delivering services. Thinking, planning, analytics, security is key areas.

Speaker 1: And number three.

Speaker 2: Number three is where it’s heading. Even more focus on security, smart mobility, advanced surveillance, and sophisticated urban modeling, and maybe soon generative AI applications too.

Speaker 1: And this is a big, but the crucial discussions around ethics, public involvement, and who’s responsible when things go wrong, those are lagging behind in the research.

Speaker 2: That’s the surprising and perhaps worrying, finding the technology is racing ahead of the conversation about its responsible use in this specific context.

Speaker 1: So. A final thought to leave everyone with beyond the cool tech and the potential efficiencies, what fundamental changes might this reliance on AI trigger in how our cities are run, how decisions are made, and ultimately how we live in them? And maybe even more importantly, what role do you have in shaping that future? Definitely something worth pondering.

[music]

What is the future for cities podcast?

---

Episode and transcript generated with ⁠⁠Descript⁠⁠ assistance (⁠⁠affiliate link⁠⁠).