Ignorance is bliss, they say, but not when it comes to artificial intelligence. Speaking at The Urban Developer ’s Urbanity-25 conference on the Gold Coast last week, Urbis director and placemaker Kate Meyrick said that while AI is treated with suspicion, its growth continues unabated. “There is so much information but so little time,” Meyrick says.  “We’re being asked to make harder and harder decisions at a faster pace and to make a great investment to find solutions for them.” In masterplanning, this has led to the likes of “city brains”, she says, which use big data and artificial intelligence to optimise urban environments.  AI masterplanning While AI was the keyword for many of the panels at this year’s Urbanity-25, effective application remains a key issue. We need to ensure that “the thing in my pocket or on my computer can think alongside me, [and] can be a genuine partner and a collaborator,” Meyrick says.  “That’s where we start to get a lot more wobbly in terms of our uptake and our understanding—we don’t really know how to apply this or where to start.” But masterplanning is the perfect nexus of science and imagination, she says, and thus an ideal place to apply AI.  ▲ Urbis director and placemaker Kate Meyrick during her Urbanity-25 presentation last week. “[AI] can iterate and evaluate so many more solutions than an architect could ever do in the timeframe,” Meyrick says.  “It can reconcile multiple competing agendas that could take me months of my life.” But picking the right “agent” is key, she says.  “All agents are not created equal” and functionality depends on the AI agent used, whether it is ChatGPT or Perplexity.  Meyrick demonstrated her own ChatGPT agent, affectionately, and a little unnervingly, named Jeffrey.  “He can look at policy. He can look at data. He can summarise 50 documents and he can generate all sorts of creative options for me around land use, around density, and he can make sure that he’s actually testing my ideas against global best practice or principles that are acknowledged as leading the world,” she says. ‘City brains’ AI is already in action in cities across the world including Singapore, Shenzhen, San Francisco and Hangzhou as “city brains”; representing large-scale AI implementation in urban environments.  Singapore developed its first national AI strategy 10 years before ChatGPT launched.  “They built a very high-resolution digital twin. It can take real-time, actual data and trend data, and model and simulate in real time,” Meyrick says.  In the Punggol district, urban AI is embedded to optimise waste, recycling and emissions activities and services.  “They’re starting to use AI to drive some of their social services. So you can get a speeding fine from a virtual speeding cop that virtually gives it to you, and they’re 100 per cent accurate.  ▲ Helsinki ’ s Smart City initiatives include autonomous buses and drone delivery of medical supplies. “You can also be prescribed by your GP equivalent a mental health chatbot.”  In Helsinki, 400 of its 800 council services have been made available through AI. “They made all the data that goes into their AI models completely transparent and evident. They talk to people to educate them about how to use it, what to expect, and all of that is available for you to look at online.” That’s not to say it’s all smooth sailing for urban AI integration, she acknowledged, citing an AI-powered chatbot that would provide New Yorkers with information on starting and operating a business in the city, that not only gave inaccurate but also illegal information.  Obviously there are pitfalls and our eyes need to be open while we adapt it into our lives and cities.  --> Fear will hold us back The challenge with AI is that it’s often not transparent enough. “We don’t actually know what data has gone into that machine,”  Meyrick says.  “We don’t know what inferences it’s actually drawn. We don’t know what it’s thinking, because it’s actually not sentient, and that’s the beauty of it. It’s intelligent, but it’s not sentient.” In the real world, it has led to instances of bias in AI systems, wrongful arrests and potential for discrimination against certain groups, with the likes of Google’s AI and OpenAI being accused of racism, and then overcorrecting to the point of inaccuracy.  But these are issues that need to be collectively worked out, Meyrick says.  ▲ The formerly rural Punggol is called Singapore ’ s first truly smart district, integrating the management of buildings in a single estate system. “We’re moving into a whole uncharted territory that we as humans have not been able to keep up with.  “We impede our own learning. It’s slower than machine learning because we have psychosocial factors like fear, ignorance, prejudice, that mean that we make a decision not to engage with AI.” That being said, AI is not the answer to everything.  “AI does not understand what it is telling you,” Meyrick says.  “It is intelligent, and it can tell you things, but it doesn’t know why it’s told you, and it does not know what the implication of that actually means in a real-life setting.  ▲ Kate Meyrick during her presentation at Urbanity-25 on the Gold Coast. “Imagination and storytelling are still more powerful than facts, and data doesn’t always imply wisdom if the operator doesn’t understand the data.” But while the court of public opinion on AI is adjourned, she says, it’s something the development industry needs to get its head around. “We really do need to understand what tools are available to us, or we’re going to find that we get increasingly left behind to those professionals, those cities, those nations or those industries that have bothered to make more of an investment of their personal and professional time or their money [in AI],” Meyrick says.  “We’re holding ourselves back by not engaging with it. But if you do engage, pick the right agent for the mission, because if you don’t, you’ll be disappointed or you’ll fail. “Don’t expect it to be able to synthesise soul because it can’t, and don’t expect to make a dumb human smart, because it doesn’t.”
