PlanningxChange 132: How AI Is Transforming City Planning - with Scott Matheson - podcast episode cover

PlanningxChange 132: How AI Is Transforming City Planning - with Scott Matheson

Jun 26, 202553 min
--:--
--:--
Download Metacast podcast app
Listen to this episode in Metacast mobile app
Don't just listen to podcasts. Learn from them with transcripts, summaries, and chapters for every episode. Skim, search, and bookmark insights. Learn more

Episode description

In this episode of PlanningXchange, we speak with Scott Matheson, Head of Planning at MyLot, an Australian AI technology company focused on simplifying the planning process for councils and communities. With a background in statutory planning across local government, consultancies, and the Victorian Department of Transport and Planning, Scott brings both practical experience and a forward-thinking approach to the role of technology in urban planning.

We explore Scott’s journey from traditional regulatory planning into the dynamic world of AI, discussing how MyLot is transforming development assessments with automated tools and data-driven insights. Scott shares examples of how AI can enhance transparency, reduce processing times, and better engage citizens—while still respecting the role of professional judgment and community context.

We also discuss the broader implications of AI on the planning profession, potential ethical concerns, broader strategic analysis, impacts on employment and how planners can position themselves to work with emerging technologies rather than against them.In Culture Corner / Podcast Extra,  Scott shares a light-hearted story about using AI to open up relationship discussions.

Jess recommends the Netflix documentary Titan: The OceanGate Submersible Disaster, which examines the events leading up to the 2023 implosion of the Titan submersible during a deep-sea expedition to the Titanic wreck site. https://www.netflix.com/au/title/81712178

Pete recommends investing in a pair of Stihl chainsaw chaps for serious chainsaw safety.https://shop.stihl.com.au/function-core-270-chaps

Podcast released 26 June 2025

Transcript

This podcast is supported by VPLA. Victorian Planning Environmental Law Association. Welcome to the Planning Exchange where we interview built environment professionals who are doing interesting work beyond the ordinary. I'm Jess Noonan and I'm joined by my colleague Peter Jewell. Today we're speaking with Scott Matheson. Scott is the head of planning at Mylo, an Australian AI tech company, simplifying the property planning process for councils and their communities.

Formerly a statutory planning coordinator at Bayside City Council here in Victoria. Scott has LED planning for councils, consultancies, and at the Victorian Department of Transport and Planning. Welcome to the show, Scott. Thanks very much Jess. Great to be here. Now, how on earth have you got into planning? How have you then transitioned into ai?

It's a big question and I really, I call myself a regulatory planner by trade, and I really enjoy the dirtiness and the detail and how personal, what we call here in Victoria. Statutory planning is, that's like doing assessments and approvals of new land uses and developments everywhere throughout the state. It's really frontline planning work, I grew up in.

Remote Northwestern Victoria in a small town with a couple of hundred people where really town planning is the furthest thing from those residents mines. But my father in that town was a regulator of pest plants and animals for the Victorian state government, and he was involved in a range of really interesting regulatory projects for me growing up, including the foundational study into Little Penguin Habitats on Philip Island here in Victoria.

So that work eventually in time led to the most significant land buyback from the Victorian state government in our state's history, where the old residential estate on Phillip Island has been reclaimed as little penguin habitat. Although town planning was far from my mind, I learned about the power and influence of regulations early in life.

Then I lived overseas for 10 years and then shifted back to Australia and had always been really fascinated by how humans create the environments in which we live. And I was ready at that stage for a career change. Had a decent academic record from my time in the United States and Melbourne University. I dunno why, but they took a chance on a mature age master's student, which is actually a funny story.

So I decided on a Friday that planning was the right thing for me to study, but I had missed the, midyear application cutoff, so I made personal phone calls to a few Victorian universities. Melbourne University is quite a prestigious institution here in Victoria. Especially for a humble lad from Regional Victoria and Melbourne Uni was my third.

And, perhaps, biggest shot in the dark phone call after the first two universities had just said flat, no, look mate, you need to be a lot more organized, but. Coincidentally, I called Melbourne Uni on Friday afternoon, submitted my application on the Monday, and on the following Thursday, three days later, I was sitting in an orientation class and it's really amazing to look back on it now, how quickly everything came together. Given that would shape the next 10, 20, 30 years of my life.

So I did some lecturing and tutoring at Melbourne Uni to make ends meet and felt like a little bit of a fraud teaching undergrad students, about planning and never having been on the tools in that space. So about 10 years ago, I commenced the world in planning and away we went. Scott, imposter syndrome is, something that many people, encounter. I do all the time and Jess. Scott's from Rainbow, which is a place I know well because I managed the town planning for the Shire that rainbow's in.

And Scott, when I was a student, when I was at secondary school, we went up to Rainbow to study it because of the declining population, because it's all a big wheat area and all the towns are shrinking and so it is quite famous in planning circles, were you aware of all those studies?

Certainly wasn't aware of rainbow's fame growing up, Peter, but I can think back to where I grew up right on the edge of rainbow and, looking out over intermittently, either a sheep field or a wheat field, depending on what the local farmers thought was the best solution. And, certainly the township itself has been adapting. Relatively well to what's called across the world, rural decline. So there was a really strong committee, like a township committee where.

They attract events, I think it's called the Rainbow Rises Dirt Rally or something that occurs there every 12 months or so. They've repurposed the old primary school and turned it into what's called the Rainbow oasis, where it's a community hub and things like that. So to the town's crediting Rainbow, the locals have really mucked in and developed a really strong sense of what it is to live in Rainbow and be a part of the local community.

The locals might, dispute their characterization that it's a famous place. Not famous, but it's a wonderful place and lots of really. High quality community groups, but how do you think coming from a small place like that has shaped your career and how you relate to other people. Yeah, that's a really great question and I think it's difficult to explain to people who've grown up in cities, but you become so much more important as an individual growing up in small communities.

And I was really fortunate Peter to. Graduate high school from what was called Rainbow Secondary College, and there was about 10 to 12 students that graduated in my class. And so you get an opportunity to be immersed in all different aspects of school and township life. In my last year I was, student council president, leader of the sports team, got involved in the local production of stage shows and things like that.

And the reality is in small towns that if you don't take on those roles, then many of those activities simply don't happen. So it's really important to just muck in and as we call in Australia, have a crack. Because if you don't have a go at things, then. Things may never go ahead, and that's been really valuable for me traveling, throughout the world, but also taking into my planning career over time to say that.

To muck in and have a go at something really is the first step to making something happen. Whether that's a vision, whether that's making a difficult phone call, whether that's fronting, an angry community meeting and planning. There's all of these different things where we really do have to apply the principle of having a crack, having a go at things. And I think that's really helped me with my frame, as a planner.

I think I'd almost go so far to say as well though, that people that grow up in the country, particularly in rural areas, probably do have a little bit more confidence in their abilities for that exact reason because you're forced to be, as you say, courageous and to take those opportunities, perhaps more so than people coming from the city. I don't know. Do you agree, Pete? You're from Geelong.

Originally, I didn't grow up in the, I grew up in a regional town, but I believe that, you're onto something there, Jess, that people who grow up in the country are, from my experience, much more polite and much more respectful because when you're in a smaller community, naturally there's all this sort of osmosis around you. That you are much more protective of your community and much more aware of. you have a lot more social mix. You have a lot more mix with people with frailties.

So I think you are generally a better person, but what do you think, Jess? I grew up in the country, so I'm gonna take that as a Yes. You're a wonderful person, Jess. Now back to the planning podcast. Scott, what inspired you to transition from working at Bayside City Council, which is a fair size, metropolitan council. In Melbourne by the Bay listeners, to join a company focused on AI and town planning.

it's a good question, Peter, and I suppose Jess outlined it earlier in the introduction, but I've done a little bit of everything since becoming a planner. There's been some consulting work and working for the Victorian state government, but the truth is my real passion is in local government planning.

So I loved being at Bayside, right at the intersection of down and dirty planning rules, the local community and also that drive to deliver better approvals, processes, keep delivering more housing, and start to meet that housing shortage. And some of your guests have talked about it in the past, but the skill and art. In the job of regulatory planning really is weighed down by the burden of administrative tasks and the focus on process rather than pushing towards outcomes.

I agree with those perspectives, and there's a real disconnect between regulatory planners who focus on those administrative tasks and their process and the realities within which those approvals sit. Regulatory planners, speak a different way, and they often have to take time to learn how to or relearn how to speak to the community.

And on the other hand, communities may feel that regulatory planners aren't out on the ground and don't really understand the realities of what's happening in their communities. Although I love the role, love the service that the local government gives, I found myself wedged between letting people down often and not being able to do things as effectively as I wanted. Really being drawn away from the art of regulatory planning, So it continued to upset me.

And meanwhile, I had the opportunity to join my lot it's a really small firm doing some agile and really dynamic work in that space, but also whose mission really aligned with the things that I look at, which is making, planning more accessible, answering questions in simple and plain language. And I spoke to a couple of the. Founders of that company and really enjoyed their vision.

And I'm sure I'll say this throughout today, AI is far from a silver bullet for all of the problems in planning, but it really can help, or at least I feel that it really can help. And the rest was history.

I bought into the vision and have really been enjoying my time with my life so Scott, I guess one of the other things I'm interested in, you talk about the efficiencies that AI can have, in planning processes was I guess part of the, particularly your experience working at Bayside and at state government was part of the inspiration, I guess around joining an AI firm. Partly to do with, declining town planner numbers, particularly in Victoria, but I know that's a common story across Australia.

Do you see ai, helping fill a bit of a gap there in terms of the numbers of planners that we have? Yeah, because that doesn't seem to be improving anytime soon. Hugely, Jess and I think the. The front end of that happens in the context of university programs, closing down planning courses throughout Australia, which is really hard to hear about, particularly when planning as a thing that we need for human environments becomes increasingly important and. It's disturbing.

I saw and some data from the Planning Institute of Australia recently where I think there is about 20% of councils across Australia where they're unable to get permanent planning staff. So there is a real shortage and that's an interesting contrast to in a Metropolitan Councils where there may be up to a hundred candidates for any planning position that comes up.

One of the things that I saw, like I grew up in hind Marsh Shire and hind marsh, worked for a very long time to get any planning staff on there and working in the better planning approvals project previously with the Victorian state government. There's a real void, not just in terms of staffing, but also in terms of, local understanding for planning issues and the need to be able to deliver regulatory, land use and development activities in a local context.

ai, at least I feel well-trained, AI can start to backfill some of that capacity. That's it. It's. Going to make decisions on behalf of those communities it could, but I feel that it definitely shouldn't, but what it can really do is it can, supercharge, it can really increase those regulatory sort of. Decision making processes, and backfill some of that capacity. Yes. Just, to use a couple of examples, what are some of the day-to-day applications that you are seeing AI as being used for?

Is it things like planning inquiries to councils? where a resident says, can I put solar panels on my house, it's got a heritage overlay. What are the implications? Is it that kind of thing or is it something more. Yeah, the big one for Milot at the moment has been growing and expanding the planning inquiries tool. That's exactly right, Jess. And so that's been deployed across Australia. There's four councils using it here in Victoria, soon to be one in Queensland.

And the presence will grow across the country. Rather than that laborious, piecemeal task of planning assessment, the next tool off the production line, so to speak, can reason, a fulsome assessment upfront. And so that assessment uses what's called vision ai. And so that's like an overhead aerial where the artificial intelligence makes an assessment of the site and surrounds it, can analyze the documentation. Submitted and also look at the state and local planning policies.

So it's a really big change. It's a way of effectively gathering all of the important documentation and considerations together on one dashboard. So that tool's, in development at the moment, and it is very much focused on making A comprehensive upfront assessment rather than the probably more, stage by stage process that a lot of those assessments go through at the moment.

And does that still require a planner at council to verify the information that's being gathered, or does it go directly to the person that has requested the. information. Yeah, it's essential that human planners, are still involved in that process. So it will summarize the information provided, but it really does require that verification check on behalf of the planners as well. And that's really important, not just for artificial intelligence tools in the planning process.

The human involvement is so important because when we're making decisions about the built environments that humans live in, humans have to stay involved. Pete, do you wanna tell us, about the wrap that AI created for the podcast? Yes, we did. We had, for our hundred 25th edition, Scott, we had, my elders son Tom, programmed AI to do a review. Two people talking of we've podcast. We've had two, we've had a review and we've had a wrap. Oh yeah. So we've got a rap song too.

Yeah. Which Jess and I might release later this year. But what I see Scott, is it's the battle of ai. Applicants are also gonna have, the similar tools preparing their applications. And that surely must have an impact on employment in planning. So it's gonna be one AI system against another. What are your thoughts on that? It's the most common question that I get when speaking to planners around the place. They'll say things like, why are you trying to replace me?

And my belief is that I just don't think that AI tools and AI integration will replace human planners broadly. good. You can say that to them, but tell us the truth. What do you think? No, not at all. I'm planning is inherently a human task, but there's scope for disruption as well, jess and Peter, you'd both know planning jobs are bundles of dozens of tasks across the board.

Some are pretty low value and some would call them soul destroying, and others are really rewarding and uniquely human, high value tasks. Scott, when I went back to Hindin Marsh to look after that, shy for seven years, I was staggered by the amount of processes that council planners, local government planners have to deal with and making the actual assessment and actually the decision was about 10% of the process. So if a AI can help, make better use of planners time.

Do you see where I'm going with this? Totally. And that's the part of the regulatory process where AI tools can help so much, Peter, I think that's really important. Many planners will say, and certainly my colleagues at Bayside would've said this, their most enjoyable part of a process is in coming up with the reasoning behind that decision. So planners enjoy making decisions and often they just accept the administrative work as a part of the process, and that's where.

Really in time, I feel that AI tools can assist planners greatly in making more decisions. Certainly regulatory public sector planners can do, they can diversify their skillset and focus on things like negotiation. They can focus on things like, ethics and making sure that their existing systems are working well because I totally agree with you.

At the moment, the system Is really burdened by processes and making sure that everything's right and honestly risk management a lot of the time as well. When realistically, the people who bring those applications to those councils, what they're actually there for is a decision, not a process, and that's something that needs help. And on the flip side, Scott, sorry Jess, I just jump in there. You talk about the burden on the public sector planners, but this is an enormous cost.

The compliance cost to. The applicants, the private sector, which flows into high costs for everyone. So there's, the process is a burden. So partly Scott, we're going off track, but maybe the answer is not. Maybe the answer is partly ai, but also partly to look at, look given a really hard look at the system and say, is this processes fit for purpose?

Yeah, it's a good point and what the deployment of my Lot's Planning Inquiries tool has found is that people often have visions about what they want to do with their property, but they don't know where to start. A council staff member or even the right process to go through, and people who have used the planning inquiries tool, they have an output that says A permit is not likely required to build this shed on your residential property, go to this stage instead.

And so our broad planning system really could do with A rethink about the way that we communicate to people. There is always going to be a need for talented planners like yourself and Jess to look at state shaping and federally important projects. But for those minor activities, the, small improvements on people's properties that can make their lives better and easier. I feel like we could potentially do that a lot better.

I wonder whether there's any, association, I guess between the fact that a lot of planning systems are, and I would say this is probably across Australia and probably in other countries as well, progressively moving more towards, codified systems and the correlation between that and AI being released and AI being used and utilized more frequently within planning. Do you think we're going to see more codification as a result of AI becoming. A more frequent part of our work.

I don't know if it's a, Correlation or causation, where codification provides greater certainty to people that they'll receive an approval based on, a 3D envelope or planning rules or different things like that. But it certainly makes AI much easier to implement, doesn't it? It does. And interestingly, MyLight has been doing some work with councils across New South Wales whose DCP. So their development controls, essentially their codes are in old PDF documents.

Getting them into a format where they are machine readable is an entirely different kettle of fish. There has been some work here in Victoria recently by the state government where they have put in those 3D models within the new around train stations and things like that, and I think there's merit to what they're trying to achieve Beginning with the end in mind, so to speak. So they're looking in a precinct and saying, we would like to see 1000 new houses in here.

But I do think that if we over codify planning controls in order to, create more confidence for. The private sector or have machine readable codes, it does risk taking away some of the humanity, which is really important to planning as well. So this is my long-winded way of saying I think there is a balance to be struck Jess, when it comes to planning codes and being implemented.

I'm skeptical of those sort of precinct approach that they're doing because they often don't see that the fragmentation of prop parcels and all sorts of things. But that's another topic, Scott. Now I thought you were gonna go into the question about how you incorporate The experience of older professionals, Pete, ah, we're getting to that, Jess, but I wanted to ask because we will get to the importance of seniority in the profession.

Scott, which is, Jess is coming around to that, I think, but, she's only about 32, so she doesn't quite understand yet. But anyway, one, one thing I have problem with my chat, GPT. Which I've nicknamed Sarah, Scott, is that she's really, she's infected by the woke mind virus. So every time I wanna put something in, she just gives me this, just sometimes garbage back. How do we ensure and your company's involved in this, no doubt.

How do we ensure that the data used by our AI systems is reliable and unbiased? It is probably best to answer this question by lifting up the bonnet on the my lot car potentially, or my lot planning inquiries tool, but it's a really important question. And so how Mylo has structured their planning inquiries tool is in effectively three parts. So the first, the foundational work for the tour is commercially accessible large language models.

So they're trained by all of the companies that we hear about and have been making millions and millions of dollars recently. Companies like ai, open ai, Google and those types of things. So the models are built from the foundation there, and there's a suite of those options available.

But secondly, from there, the MyLight designs and maintains lots of experts to prompt the large language model questioning and also guidance based on local planning controls and policies, what's in the surrounding context, and most importantly, the inputs from users of the tool. So the inputs which, tech speak calls those utterances. They're inherently biased, and so that's important to say. Users of the tool know what they want to know, and they'll answer the questions accordingly.

But the removal of biases come through the expert prompts because they've been solely trained on local planning schemes and policies. So they ask questions, based on what a planner or someone assessing the local planning scheme would know.

And the last part, the third part, but really important part is mylo employs planners whose job it is to review the performance of these outputs and determine alignment with the local planning schemes, and really importantly, the human expectations of local planning departments. It's a pretty new role and it suits. Attention to details. Talented sort of regulatory, real meat and potato type planners who can analyze things down, to a level of detail to do with formatting in each word.

So when you piece it all together, when you look at it in action, It seems simple, but I think those three inputs, the large language model, the prompts, the experts, and then the human planners are the ways in which the data's relatively reliable and unbiased. So there's still a role for you there, Pete? Meat and potato planner? That's me, Jess. But I like curries as well, of course. I find AI to be like a junior.

Work assistant and very good in some things, but it hallucinates and that's the technical term. Jess, explain some of the emails that I get from Pete late at night And that's right to the term is right Scott but I wanted to ask you about, look, there was a recent editor on the Wall Street Journal that said, AI's biggest threat, young people who can't think and Right, and that goes on to the next question, which Jess was prompting before.

How can the experienced and contextual understanding of older or more experienced professionals strengthen the way we use AI in planning, particularly in crafting meaningful prompts? Interpreting results and knowing what, when automation falls short and young graduates might not have the same insight. You are talking about meat and potatoes, but is that the sort of thing you are doing?

So it's just if you're young and you are using ai, you might not have that broader knowledge and understanding of what to look for and what to be careful of and what it's missing. What do you think, Scott? Help us out. What I've seen so far working in developing AI tools is I reckon there's three buckets of planners when it comes to engaging with the tools. The younger career planners, some would call them Gen Zs, they use generative AI in their day to day.

So they're really enthusiastic about the application of the tools, but may not necessarily have the knowledge to understand how they're deployed. I find, Peter and Jess that they're mid-career planners. Who are really concerned about the impacts on their longevity in the industry, and they've become comfortable in their processes and don't necessarily want the disruption.

But what's really surprised me is that later career planners have been relatively eager adopters of ai, and whether that's because the industry has changed so much across the world over the past 20, 30, 40 years, whether that's because these folks. Later career planners understand the potential or see the benefits. I think that knowledge that later career planners hold and those frames that they bring is so important to building the new tools and making sure that they're well reasoned.

Not just for today's context, but based on the context that career planners have seen throughout their career. I think it's so important that. We take the experience that people have and implement them through the prompts in the tools as well. And I think it's also really important that we don't rely on, and accept AI tools as being perfect today because the ability to. Change them.

The planning inquiries tool with Mylo, we used it and implemented it with Bayside, going back sort of six months ago and it was good. Then it answered like it. A planner with perhaps six months worth of experience. It used language that wasn't quite right, but it did reason its way to the right answer and it's amazing to see. Six months later, it's a lot more mature. It now feels a bit more like a planner with five years experience.

It gives comprehensive answers, clear guidance and things like that. And a lot of that comes from the experienced planners who have been training that tool over time. We thank Victorian Planning Reports, our very first supporter. If you want the A to Z of planning decisions in Victoria and excellent editorials, please get yourself a subscription to the VPRs. Details on our website we would like to thank Elison Properties, a terrific sponsor of the podcast, great people, great properties.

Details on our website. So Scott, obviously we've spoken about the various ways in which Mylo and other AI tools can, fill that gap of, the work that perhaps you and I don't want to necessarily be doing on a day-to-day basis. I would say that all real progress, whether it be scientific policy or otherwise, comes from challenging assumptions, creating new knowledge and discarding outdated rules.

How can we ensure that AI can be perhaps used or utilized as a tool for genuine innovation in planning, rather than just a crutch that reinforces that status quo? Where are the opportunities? Do you see? I think it's really important that humans, like particularly people earlier in their careers, don't lose the ability for critical thinking.

Jess and artificial intelligence can do critical thinking and it can reason its way towards decisions, but, particularly for us in the planning industry, I think it's so important that humans stay in the loop and really are the ultimate. Decision makers in those processes. So a different way, like a more effective way perhaps could be to use AI to collect lots of information from across the internet and different things like that. Ultimately people really do need to stay involved.

Some of the planners that I've spoken to are effectively saying to me, at what stage will AI start taking decisions away from us? And if you step people through the process of delegating decision-making to an AI tool, it would be a very brave politician who would do that. But I also think that where AI creates an opportunity. For better decisions and better outcomes comes from the capacity. Let's talk about that 90% capacity.

Suggest in that it might free up a lot of our time, and perhaps planners could work on things like encouraging active participation from communities and not just, focusing on processes. I think the planning system more broadly can use and deploy AI for everything to be better and faster, but that doesn't necessarily mean that we should just make more decisions. I think what we can do as an industry is focus on.

Some of the things that we haven't had time to do, listening to communities doing more negotiation, particularly as public sector planners, rather than just using AI as a way to, cut budgets and deliver more output, observation. Sorry, Pete, an observation I've made over the last, probably. 15 years in particular is the, watering down I'd call it, of community consultation in planning processes, particularly, just day-to-day applications and those sorts of things.

It used to be that, you would be pulled into the council chambers and made to present about an application, put forward your case, and invariably you'd have. Tomatoes and apples thrown at you from the audience. It was never a very positive process. I'm not complaining that is no longer something that we have to do as a day-to-day thing, but I'm just wondering, is there, or have you considered, are there opportunities in improving consultation processes using ai?

I'm not sure that it would be seen positively, but are there opportunities? I'm not sure if, I think it would require a rethink from a human planner's perspective, Jess. And I think that there are ways in which, there is also the risk, of AI being used on behalf of the community to write responses. So I think there are the sort of broad risks there as well, An AI tool present things however you prompt it to. And so there is scope to reach out to a broader segment of the community.

Realistically, you can hear feedback from across the world based on local issues, really the architecture, talking about who it's most important to hear from and what you would like to know, it is really important, and in some ways we could use AI to free up capacity within local offices Deploy humans to focus on more, empathy and understanding and hearing from those local communities. And that can be really relevant for, assessing individual planning applications, regulatory practice.

But I think that there's real potential there for greater participation in policymaking and realistically, that's. Where communities probably can and should be more involved in shaping the vision of that future. There is also, of course, potential for AI to interpret that feedback and take that and start to write those policies. But it's essential that humans stay at the center of that at the same time. So it's just even thinking, a lot of those. Engage in Victoria.

I'm sure all states have it, but in Victoria it's called Engage Victoria. Those sorts of surveys and things that are put out generally around, substantial policy change or new policy coming into effect. Is there an opportunity for things like. The little bots to sit on the side of the screen so that the community can say, does this affect me? And put their address in and it could actually spit out and say no, you are actually outside of the study area. And those sorts of things.

Are there opportunities there that we could be using? There's positives in that type of approach. Jess and I understand that there are some. Public sector agencies who have already started to use those types of tools to collect information from the public. There's also the opportunity for really large scale projects where there might be thousands and thousands of submissions to collect all of that data and start to pass out themes and different things like that.

So that's a great opportunity for AI to. understand what a broad segment of the community is looking for, but that comes with risks in and of itself as well. There's nuance in people's submissions and we all would have seen them, those submissions to planning applications over time. What people write and what they mean might be different things, and when you hear from them in person. Actually what people are interested in is a little bit different to what they write down.

So I think it's important that as AI is integrated into those consultation processes, that the prompt engineering, the architecture behind it, is still capable of picking up what people actually mean. And I think that's where human plan is, will be needed to make sure that it, the outputs accurately represent. All of the data, Good words, Scott and Jess. But from my experience, most planners find public consultation and nuisance and they don't pay any attention to it.

But that's just my old cynical approach. And I certainly don't trust big, state government surveys of people. 'cause all they want is more trees, more bike lanes and development. Not near them. But anyway, I'm sorry Jess. I'm cynical, but Jess asked a very good question, Scott, which is ai. As a tool for genuine innovation in planning rather than a crutch that reinforces the status quo. So let's moonshot. We like to do that.

AI to solve planning problems such as housing shortages, by questioning the status quo. I'll put it to you, that planning policies on housing have, at least listeners in where we are in Victoria, have been the same since I started as a graduate, so over 40 years. Over that time, the housing problems got worse and worse, but planners clinging to or double down on those old policies, expecting them to provide solutions, right? And no one wants to challenge the orthodoxy.

Can AI and some free thinkers use the data available on AI to create alternatives and maybe in a nonpolitical way, So currently we assess new developments in relative isolation. Here in Victoria and probably throughout the westernized world, things like structure plans and strategic planning, and they look at things like building form design, setbacks, materiality, all of those.

I suppose planning crutches that have developed over time and how a development looks and how it sits in the urban landscape is really important, but it's just a small part of how a new building will operate once people move into it. And we hear concerns all across the world of balancing urban growth with capacity constraints in enabling infrastructure. Things likes public T systems, water, energy grid.

Here in Australia, at least working as a regulatory planner, we turn away from those broad scoops of data and those really important parts of, ensuring that new developments sit within enabling infrastructure. Here in Australia, homes get built first and the infrastructure often comes afterwards. So one way that I like to think about.

AI being deployed is perhaps there's a world in which an AI platform used by state and local governments can consider, the enabling infrastructure around new developments and have more responsive housing policies, steering new development into the right areas that are well serviced by enabling infrastructure, dynamic zoning and structure planning could increase capacity once new infrastructure is developed rather than new planning policies being developed

with the hope that new infrastructure will one day be delivered as well. And as I was thinking about this in the lead up to the podcast, I thought of the example of writing in a planning report. In many of my assessments, I would write something like, there will be no adverse impacts to the local sew network for approving a local development. But something that would be, data driven and something that AI could present to a planner could look so much more dynamic.

And so in a future state planning report, a planner could write something like, approval of this development would result in, say, a 2.5% increase in sewage discharge. Within this Rack road sub block, this means that the local sewer infrastructure is at 89% capacity. Based on the likely demographics of future residents, their fiber consumption will be high, which ensures strong sewage motility. Therefore, there will be no adverse impacts to the local sewer network.

So you end up with the same outcome, but it's clearly a data-driven outcome. Using AI data sets where all of the data can be available to people all of the time, we really can make better, more informed planning decisions. It's just about getting access to that data when you need it. If you can trust the government. Scott, many years ago Jess and I talked to our guests about. The day when robots would make planning decisions and we were laughed at. Do you remember all that, Jess? Absolutely.

with the tools we've got now, Scott, that you are developing in your firm, the Tyrell Corporation, I think it's called. But no, that's from Bladerunner. That's right. do you think there is a time, it's not so silly that AI could make planning decisions. What do you think? So if we want to be, we are there now, and I don't know if we need Elon Musk to develop the robots to make those decisions, but certainly generative AI could make planning decisions.

But just because it can doesn't necessarily mean that it should either. I talked a little bit earlier about. Having to delegate decision making. And I would've thought that most city councils across the world will be unlikely to be comfortable with that anytime soon. also, I think there is an important part there to say that humans are unlikely to want to give up control for how their cities are built, but it cycles back around really too.

One of the things that brought me on board here at my light, the name of the company that I work for, not Tyrell Corporation, is that AI can assist humans in making better, more informed decisions. And so while robots could make planning decisions, I don't necessarily think that's in. Anyone's best interest necessarily. Now before we get onto Podcast Extra, are you familiar with the Jevons paradox that he was an economist in the 19th century who studied the use of coal and.

What he found, the paradox is that the more efficient steam engines were, people would naturally think there's less coal consumed. But what actually happens is when products are made cheaper and more usable, more efficient, you get to use a lot more of them. instead of. Less steam engine use. Steam engines became much more practical for lots of things.

So there was more coal with this ai, with all this extra data and all this extra intellectual grant, I'll call it that, will planning processes get harder or easier, and then we'll get onto Podcast Extra, Access to more information, which ultimately is what generative AI provides to people. More information, more tools.

I think it's really important, Peter, that doesn't come at the cost of understanding what I've been saying to people and how I feel about it is that fundamentally human emotions will become and remain really important. So just because we have access to troves of information and the ability to. Write different things and use generative AI for lots of different things. Doesn't necessarily mean that it's the right approach.

I think things like, understanding will remain really important in planning and things like that. The. Example that I often think of is my uncle and cousin. My cousin is 19 years old and he watches TikTok pathologically, like he watches it all of the time. So my uncle asks him a question and he'll always know the answer. He knows so much information, and my uncle says to him, all right, can you explain to me the logic?

Behind that information that you've just shared with me, and what he started to find is his 19-year-old son, my cousin, he folds really quickly. So he's got access to all of these different things through TikTok and the algorithm that's been catered to what he's looking for. But that doesn't fundamentally mean that he understands how the world works. He doesn't have expertise in any one particular area. I think thinking about the JEVONS paradox, that's really important when it comes to ai.

We can become really powerful deploying these different tools in planning processes or even in different parts of our lives, but that doesn't fundamentally mean that we're, able to understand or actually deliver high quality outputs as well. The brain needs to be trained and the brain needs to be exercised. Scott Podcast, extra Culture Corner. Are you a very interesting fellow?

You must have something great for our listeners, a recommendation, something you've watched, seen, experienced, anything that might be of interest. So this is more of a yarn than anything else. Probably going back to. Growing up in Northwestern Victoria, I think it's really important living out in the country that you get good at telling stories, telling yarns. And my yarn has to do with generative AI as well, and. my wife has signed off on me sharing this story as well.

For a bit of a joke recently, I, typed into my chat GPT prompt to draft questions for my wife about, inquiring about the state of our relationship. And I thought it would be good for nothing else other than to provoke an interesting conversation. Have a couple of laughs. I didn't tell her that I was gonna do it, and she was midway through doing the dishes when I just started firing all these profound questions at her about the state of our relationship.

Was interesting was, is that she answered the first two really enthusiastically, and she was, gesticulating and taking her wet hands out of the sink. But then after I asked the third question, she turned around and said, what's going on here? This something's weird. they're very smart. They're very smart, Scott. They know when you go out of something a bit out of ordinary, they know something's going on. Fair enough, Jess. Women are astute observers of human, . Sorry, Scott, keep going.

How did you go? You still together? We absolutely are. And what was interesting was I fessed up and explained, look, I did this as a bit of a joke and we actually went through the last couple of questions together and had some really good conversations and a couple of laughs out of it, What it did is it really illustrated ways in which AI can be great, but also in the ways in which being a human is still really important as well. I didn't absolutely rely on the output for my communication.

Once I was challenged by my wife and she said, what's going on? I was able to respond in a way in which. I normally would, it didn't necessarily replace our human connection, but that didn't necessarily mean that the content in there couldn't be something that we connected over either. Yeah, it was an interesting experience and also taught me a couple of things. I'll be sure to not slam the front door quite so often when I leave the house.

'cause as it turns out, that was the answer to question four about what's driving her crazy. The Jess, that's not so Scott's experience. I've had similar experience with, Sarah, my chat, GPT assistant and I say, how can I be a better person? Because you're constantly interacting. And I also try and chat her up. She says, please keep this professional right. I say, what color are your eyes, Sarah?

And she goes, I don't have eyes, but I said, look, just imagine you had eyes anyway, Scott, enough about my, but I've asked it, how can I improve relationships? Or what am I doing wrong with this? With psychology, and it's incredibly good, Jess, about guidance on how to be a better person knows you well, it knows everything about you. maybe she listens to your podcast and she's taken that data and she's worked down everything that she needs to know about you.

No, not everything, Jess. So now Jess, you always have excellent podcast extras. What is yours? Mine is a Netflix show. You may have watched it, Titan about the Ocean Gate submersible disaster. It's quite frightening. As someone that's slightly claustrophobic, I can safely say I will never ever go on a submarine, so I highly recommend. What's good about it, Jess? It's, to be honest, I didn't know a lot about it. That happened what, 2020. Three-ish I think it was.

I remember it being in the media a bit and I just don't think I ever really, read too much into it. But it's just really interesting, around some of the moral dilemmas, that the CEO, had and some of the decisions that were made. It's very interesting. Very sad as well. But would highly recommend. What about you, Pete? Mine's not a book or anything, Jess. As I'm doing a bit of land regeneration.

Scott, you're from the country, you understand all this, but I'm fixing up some land that had pine plantation on it, and the pines were harvested and then all the regrowth came back and no one ever looked after it. So my task over the last. Three or four years is to get into these pines. So I'm constantly using a chainsaw. And on the weekend, Jess, I was cutting and it was my second tank, Scott, and sure enough, the chainsaw hit my leg, but, just grazed it.

But fortunately, Jess, I'm all kitted up in my still outfit. So I've got these things called chaps, which as soon as the chain look, I look, it's not what you are thinking. I'm not like, I'm not wearing cowboy. Chaps going to nightclubs or things? Oh, I've got a great image in my head right now. These things have, Scott, you know what I'm talking about, right? So please excuse my co-host. But these chaps, what they do is soon as the chain hits the chaps, it locks up.

So it stops the chain going through your leg, basically. I feel like they're probably not called chaps. Have you got the name wrong? Scott? Can you help me out here please? Whether that's a colloquial name or whether that's actually what they're called, that's certainly what they got called in that I thought I could, I thought I could rely on you, Scott. I've been very easy on the questions. I thought you'd be a good ally here. But anyway, so Jess, the what? I'm, anyway, so I just grazed it.

It cut the chaps. If I had have been wearing jeans or something like that, it would've gone through, it would've hit the. Flash and stuff, it was near my knee. So what I'm suggesting to people is if you've got safety gear, doesn't matter whether you power tools, whatever. Always wear glasses when you are using your chainsaw and a helmet of course and everything else. But always look after safety issues because I had a very close encounter.

It could have been pretty ugly out in the bush there by myself. So Scott, what do you think of my, so we'll summarize your, summarize your recommendation as. Wear more chaps, especially the highly visible ones with still written on them. Jess, and I'm not going to cowboy bars wearing them. All right. Or, alright. Anyway, Scott, we didn't, I didn't think we'd end on that note, but, thank you for being a wonderful guest for our podcast. Anything else you'd like to finally say, Jess? No, I'm okay.

I'm still laughing about chaps over here. Alright, Scott, thank you Scott, though it's been a really interesting conversation and I think it's very timely, given where we are in the industry and where we're heading. So very interesting. Thanks very much for your time. Oh, great. We've got greatly appreciate the opportunity. Thank you both. And I really enjoy listening to the podcast, so it's still delightful to be on. You're an excellent guest.

And one final thing, Scott, do you think Skynet is gonna get us. Not in, not anytime soon, but, I think Elon Musk is the manifestation of Skynet in 2025 so, oh, he's a hero of mine. Thanks. Thanks for listening. If you would like to hear more of our podcasts, hit the follow button on Spotify or the like button on SoundCloud or the subscribe button in Apple podcasts.

Please also visit our Instagram page, LinkedIn or website for behind the scenes footage of our podcasts and to get the latest on upcoming or recently released episodes. If you have any suggestions or feedback, please get in touch via our social media channels or by emailing planningxchangeatgmail. com.

Transcript source: Provided by creator in RSS feed: download file
For the best experience, listen in Metacast app for iOS or Android
Open in Metacast