Blueprints of Tomorrow: AI's Role in Shaping Our Urban Future

Artificial Intelligence, or AI, is more than just a rapidly evolving futuristic technology. It is already an integral part of our daily lives, from the recommendations on your favourite streaming service to the algorithms that power critical business decisions. 

But what is AI exactly, and how does it work? How can we use it to create smarter, safer, and more sustainable solutions for our industry and the built environment? 

In this episode, James Mcallister, Smart Design Lead at WSP in the Middle East, is joined by Massimo Dragan, Unit Leader Digital Innovation, at WSP Italia to discuss artificial intelligence (AI) and its impact on the built environment. They explore the differences between AI and machine learning, the benefits of integrating AI into business operations, and the challenges of ensuring responsible and ethical use of AI. They also highlight how WSP has already leveraged AI tools in their processes and discuss the potential impact of AI on job displacement. The conversation concludes with a discussion on emerging trends and the future of AI in the built environment.

Want to hear more insights from leading industry experts? Subscribe to our podcast via Spotify, Apple Podcasts, and SoundCloud to stay up to date.

Transcript

Artificial intelligence or AI is more than just a rapidly evolving futuristic technology. It's already an integral part of our daily lives from the recommendations on our favorite streaming services to the algorithms that power critical business decisions. But what exactly is AI and how does it work? How can we use it to create smarter, safer and more sustainable solutions for our industry in the built environment? My name is James McAllister, Digital Design Lead at WSP in the Middle East. And in this episode,

Anticipate podcast, I'm delighted to be joined by Massimo Draghi, unit leader, digital innovation at WSP. In this episode, we'll demystify the complexities of AI, explore its nuances, applications, and the impact it has on our world. We'll also discuss some of the challenges that AI poses and how we can address them. Massimo, welcome to the Anticipate podcast. Thank you, my pleasure. Good day.

So for our listeners who may not be familiar with machine learning and artificial intelligence, can you explain the differences between these two terms and perhaps tell us a little bit in simple words how AI and how it handles structured and unstructured formats of information? For sure. So artificial intelligence is kind of a broader field that deals with creating systems or programs that are capable of exhibiting behaviors that we would define as intelligent. And the goal of artificial intelligence,

is typically to develop these systems that can perform tasks that would require a certain level of complexity. Machine learning, on the other hand, can be deemed as a specific subset of AI, which involves teaching machines to learn from data. So it's more focused on more on the analytics, on the application of statistical methodologies, and so forth. The characteristic of machine learning solutions is that

They aim to improve their performance on task over time without being explicitly programmed for it. And then on your second part of the questions on structure and unstructured information, we consider structured information as typically data that comes in a numerical and organized format, while unstructured information would be as simple as a document with sentences, concepts, so where it's more similar to the way that we express.

concepts and knowledge in a conversational manner. Organizations are increasingly adopting AI, so maybe you can share some insights into how businesses are benefiting from integrating AI into their operations and some of the things that we're seeing as success stories potentially. I think there's probably two or three key tenets in this area. We're seeing organization increasingly leverage

the support of AI tools in the ideation process. So where you can get a lot of input from many different people and then at the same time use sort of an assistant to help you validate those idea or even augment those idea with additional information that the team may not have direct access to. There's also the side where AI is kind of becoming, there's a frequently used term is copilot. So it's in a day to day.

It's like having somebody sitting beside you virtually to help you navigate in the complexity of tasks that you do on a database. So perhaps there are things that are happening that should capture your attention or that are more urgent than other ones. Or there's a flurry of information, your 2000 emails per second that you're receiving versus a lot of maybe documents that you need to look into and the likes and these solutions.

have an unparalleled ability to sift through a lot of information and synthesize that content so that they drag your attention where it's more needed and extracting the relevant concepts from potentially larger corpuses of knowledge. I think copilot's an interesting thing, as you mentioned, being an assistant to how we work, not a complete...

replacement for the tools and the people that we're talking about. And obviously how we implement this in terms of the governance and monitoring are crucial aspects of kind of AI solutions. So maybe you can elaborate on some of the important factors in ensuring the responsible and ethical use of AI. So the I think you touched on a very important point and I think Copilot is the right

is the right term. There's a lot of concern that these systems will sort of take over or replace things that we do. I think what will eventually, we should be considering them more as tools or mechanism that would allow us to perform activities that we still are in control of in a more safe, efficient, and potentially more informed way. So,

It's true governance and monitoring are two important aspects. And I'll touch on two things. One is the governance piece, I think it's paramount to a lot of the companies that are proposing and put into the market these tools for companies. And it's one thing that we use with a lot of attention. We're reviewing offerings that we see in the market and we're only partnering with providers.

that would allow us to develop these solutions for our clients, that we make sure that the information is treated the way it should be, in a safe way, that access to information is secured. And on the other side, the monitoring piece will always be leveraging the expert judgment of the end users. So in our case, we are a knowledge company, so subject matter expertise is paramount and essential to validating

that the responses that are generated or the insights that are proposed by these tools actually are grounded in scientific and technical merit. And that the information that is being synthesized, assembled, or potentially newly generated actually makes sense. And that side of working with these tools rather than having sort of the...

the magical expectations that these things by themselves will solve every problem I think is probably a way more likely future that we're travelling towards. So they're not going to take over the world just yet, then we're saying we still need the human interaction, right? There's no Terminator -esque horror story. I don't think so and you know both from an institutional perspective as well as from the industries. There's a lot of attention, emphasis on the ethical aspects of this as it should be. So there's even, you know, people will be familiar with the fact that tech companies have actually asked institution to put a good regulatory framework around this. Recently, the European Union organized a summit to tackle this. So there are...

As with any, you know, even a car can be very dangerous if you don't have a code of conduct to use those tools on streets. So this is not going to be any different. And I mean, obviously, naturally, new tools require new policies, new ways of working, because there is obviously, like you said, the opportunity for abuse of certain tools.

And I think maybe something people don't necessarily understand is how WSP has already been leveraging these AI tools. And maybe obviously with your expertise, you can explain a little bit about some of the use cases of how we've integrated AI into WSP processes and the things we're doing and things you and your team are doing. Absolutely. So there's, I think there's two aspects here and let's go back to.

that initial distinction between general artificial intelligence and then more data -driven machine learning components. So we've been using machine learning for years now, and it's been helpful to gather insights from large amounts of data and information, and has allowed us to also develop predictive models. So where we would be using insights that are present in the data.

to predict future states of, for example, the environment. So our team is currently involved in an air quality monitoring project where we're using for safeguarding the exceedances of air quality measured at construction sites. So we're using and developing predictive models that would use environmental conditions to predict what if the conditions of the...

at the working site are going to be safe. On the artificial intelligence side, we are actually developing with the use of a large language model. So people will be familiar with the, you know, the advent of ChatGPT and these natural language processing interfaces. So we're using these technologies to develop conversational tools that would allow our teams, but also.

people that are not necessarily technically versed at highly complex content to be able to interact with thousands and thousands of pages. We're currently piloting one on a larger remediation project that has been in the design and implementation phase for more than 20 years. So you can easily imagine the vast corpus of data, information, expert opinions that has been solidified.

in thousands of pages of reports because that's the way that we could typically communicate among each other. That's the way we use to explain what is happening, what our decision -making process was and what our suggestions are to complete a specific task. In this case, give back the land to communities that used to live in an uncontaminated area. So in this case, the application that we're developing,

will allow a human interaction with a larger set of information and knowledge and you'll be able to interact with this in a natural way. So by using just your language, which is I think one aspect of artificial intelligence that I find extremely interesting, the way that we'll be interacting with systems and tools in a way more natural way, rather than having to

either create code or learn how to use specific software or learn different paradigms on how we interact with information or how we extract the knowledge. The great benefit of applying these sides of artificial intelligence is that the interaction is the same as we're interacting among ourselves in this conversation. Really interesting. I think you probably touched on

probably the next point of potentially, I mean, it wouldn't be a discussion about AI without talking about the fear of someone losing their job from AI doing a better version of that job. And the fear of job displacement is obviously real to many, many, many people. However, I suppose yourself and me are more well versed into the tools and you've mentioned obviously about it being a co -pilot, something to assist us.

It's not anywhere near a space where these tools are taking over. There's always a human element that must be included into the use of AI and these tools. So what are your thoughts on this kind of concern of job displacement and how can individuals and organizations navigate the potential impact on employment? I think a better term is

that we could be using is what type of transition will the adoption of AI allow us to do? And the transition will be from current paradigms of how we accomplish our tasks, our jobs, our aspirations versus how will these assistance, these mechanisms that will augment our knowledge, these tools that will simplify a little bit the complexity that surrounds us.

How are these going to allow us to eventually make better use of our time to stop wasting time to do repetitive tasks, which is a curse that in a number of organizations, especially when you're doing knowledge intensive tasks, we all experience the deluge of information. The fact that we don't seem to be able to cope with the amount of tasks and the speed that these tasks are now required or expected to be executed.

So I really have a lot of positive aspirations and a good feeling that this is actually going to give us back a little bit of the time that has been captured by some of the way that we're currently using computers, software, and how we're interacting with the data and information to eventually make decisions. Back on the importance of the human factor,

There's various publications that predict that actually the experts are going to be essential because overseeing and validating what these tools will be generating is going to be paramount. And I think the other interesting side of this, back to the idea that the way that we interact is in a more natural way, is going to be making accessible to wider audiences.

won't necessarily require deep technical expertise. Some of those insights, some of that information is the, you know, there's that this interesting feature, one of the use cases of large language model is you ask a question and in that question you say, explain that to me as if I was a six year old baby, right? So which is interesting in the sense that I think we are, you know, increasingly becoming a society where,

technical expertise and knowledge becomes more and more sophisticated. But we also need to make sure that that's democratized, that everybody has access to that type of technical knowledge. And again, these tools and mechanism to make it simple but still understandable, I think are very, very interesting. And my expectation is that we're gonna see them adopted in various use cases.

So my take is that jobs won't go away, they're gonna change. The way our intellect will be used probably in a slightly different way and ideally and hopefully in a more meaningful way to tackle the larger challenges that our societies are going through. And we're already seeing the benefit of using these capabilities to tackle.

tasks that at the moment are daunting for us because we don't have in our brains the ability to process all that information in an efficient and rapid way. I mean, I love the idea of a kind of intelligence or data economy that these kinds of tools build for us, being able to ask a question of a thousand minds through these kind of language models. And I think there's...

There's a real interesting kind of adaptation that these can have to our work, like you've said, and just the need for the business around us to constantly find ways to optimize our own time, buy time, gain time, save time, anything. We spend a lot of money on saving time. So these tools are a natural progression, surely, as you obviously understand.

are a natural progression for how we just save time in the future. I suppose specifically around the built environment. A few years ago, everyone was putting sensors and bits into buildings to optimize the efficiency of a workforce, save time, save money on operations, maintenance, you know, there's there's time to be saved everywhere. And as you said, we are a people business. So one of our biggest expenses or the biggest expenditure we have.

is people. So if we're able to save them time and save them kind of these longer processes that could be instantly shortened by using some of these tools, I think that's a fantastic point, Massimo. So maybe kind of looking ahead to how we might continue to use these kinds of tools or the evolution of AI and how it affects us as an industry and how it will affect us in terms of what we can provide for our clients.

Are there emerging trends or developments that you find particularly interesting or rile you up into a place of excitement? I think there's a couple that come to my mind, especially for the built environment, I would say. One is a general trend that we see is that the evolution of AI will be that it will disappear from the radar. It will become in itself more of a natural thing that is out there. I think...

we've started the use of artificial intelligence with the paradigm of asking questions and receiving an answer. I think we're moving into a paradigm where it's going to be interactive experiences that are going to be more natural and engaging so that we're going to be using more senses to be able to interact with an environment that will kind of understand what we're doing and it will help us

accomplish our tasks or that will keep us safe. I think safety and the ability to protect people from events where our ability, where our senses or our ability to react may not be sufficient to prevent harm is an extremely interesting type of application where these tools deployed in a built environment may be extremely efficient. We hear about

road accidents, we hear about things that happen where you would consider yourself in a safe place and still people get hurt and harmed. The other piece I think is, as you mentioned, there's been a period of time where the built environment has been sort of loaded with senses, with digital senses. Now I think is the time where these...

sources of information with the use of AI will finally deliver on the promise of how we're going to coordinate all these sources of information and how we're going to extract the right insights, how we're going to make our places where we live more sustainable, how we're going to be able to reuse energy or waste less energy or consume less resources and the likes.

a number of these challenges, I think, reside in the ability to now coordinate all of these sources of information. And these tools, I think, will help us get there way faster. Another example, I think, is in very practical and day -to -day job. There are already applications that are used, for example, in maintenance and the services.

industries where just by talking back about the use of the senses, you need to do maintenance to an equipment that is exposing false behaviors. You can easily look at that and then you would have the interaction where you're looking because you're using devices that can see what you're seeing and can understand where you're looking at. Then the

The copilot can understand what is that piece of equipment, can gather the technical information about it, can help you do your maintenance in the right way. It can show you things that you don't see. So the use of augmented reality, for example, in a built environment where all the cables, all the equipment, all the sensors are hidden. This is huge when you're operating in an environment where you don't see things. Having an assistant that can...

feed you information about where are the cables, what are the connections between the different type of equipments and where could the source of the problem be to make your task of fixing it safer for you as a front -end operator as well as more effective so that equipment works and that you don't need to go back there. And these have all become

accessible work cases and accessible applications of this type of technology. Massimo, what's the future for AI? What's the future for AI in our business then? I think the immediate future is unlocking the knowledge capital that has been solidified and taken away from being practically usable.

We have thousands of expert knowledge insights that have been locked in documents in fragmented storage devices and the likes. I think that the first immediate use is that we'll be able to leverage the collective knowledge of the WSP subject matter experts in a way that is

that has not been possible in the past. And I think that we'll be able to help our clients leverage and extract the value from the knowledge that they have also accumulated over the years. And that somehow in the way that we've used our computer systems has sort of solidified that, but also made it way more challenging to reuse and capitalize. So I think this is probably, I would...

call it the low hanging fruit for AI. Then I think the other interesting concept is that part of this will be discovered as things happen. I think use cases will emerge on what are the areas where these assistance will help us do things more safely, do them more efficiently.

I think ultimately there's going to be not too far down in the future, the opportunity to change and challenge some of the paradigms that we're currently using. So one of the principles that WSP talks about is challenge the status quo. I think this is a fantastic time to leverage this technology to do just that. And we now have in the hands tools that will augment our...

knowledge capabilities so that we can challenge the status quo even more profoundly in, as I mentioned before, striving to solve the fundamental challenges that our societies are exposed to in their evolution over time. Massimo, thank you so much for joining me on the podcast today and for the invaluable insights you've shared. I'm really excited to see what you and the team do with it.

the future and rest assured I am not terrified that my job or anyone else's job or the robots taking over none of it you know you put me at ease so thank you so much to our audience thank you for listening all the way through and please leave a comment if today's episode has sparked your interest and don't forget to join us on the next episode.