At first glance, cities and AI are a good match, with AI having the potential to optimise many operational aspects of urban life. Cities are messy, dynamic and full of data. Luckily, AI thrives on this chaos. It’s not surprising then that technology firms are rushing to integrate AI into their urban software systems to help optimise urban traffic flows (a major contributor to greenhouse-gas emissions), reduce the power demand of street-lighting systems (1-3% of electricity globally) and model better waste-management pick-up routes. The slogan “AI for Good”, coined by the United Nations International Telecommunications Union in 2017, offers a tidy moral wrapper for this convergence, and “AI for Good” events provide plenty of examples of practical applications.
But as I’ve observed while working with public-sector digital initiatives, this courtship is more complicated than it appears. Cities are not just networks of roads, pipes and power lines waiting to be optimised; they are human and ecological habitats. Citizens want well-being: safe neighbourhoods, social connection, green spaces and economic opportunity. Nature is critical to successful cities and needs clean air, flowing water and biodiversity for systems to thrive. People and nature’s needs overlap—doctors now prescribe time in nature as therapy, and green cities attract the best talent, good jobs and great culture—but this critical element is not something that AI can address.
The resource appetite of AI
The challenge goes deeper. AI’s own appetite for energy is voracious. According to a major report released in April by the International Energy Agency, data centres are projected to consume 945 terawatt-hours of electricity annually by 2030—more than the current consumption of Japan. AI is the principal driver of this surge. In Ireland, data centres already use as much power as all urban housing combined.
Water, too, is drawn in staggering quantities to cool the vast racks of processors crunching AI models. As cities face growing water stress, the demands of digital infrastructure may compete directly with human needs and nature’s needs. In Ahmedabad, a city of 9m people in India, urban heat effects are so severe that 800 people died of heat stress in one week in May 2010 when the temperature reached 47 degrees. This tragedy resulted in India implementing its first Urban Heat Plan in 2013, now replicated across 250 regions, cities and communities in India. Under this plan, the water supply to Ahmedabad, the largest city in Gujarat, is limited to just one hour a day. Under these circumstances, the idea that precious water would be diverted to cooling data centres is difficult to justify.
Historically, cities grew where water was abundant—at river crossings, deltas and springs. My own city, Bristol, owes its name to the Saxon word Brycgstow, “the place by the bridge”. However, AI might mean that we may soon witness data infrastructure shaping planetary geography as economic and cultural lifeblood flows through fibre-optic cables and cooling towers, much as rivers traditionally shaped our urban geography.
The urban century
Cities are the original platform business model—centralised hubs of commerce, culture, innovation and governance. Their density catalysed the rise of the bourgeoisie and the modern economy. Today, more than half the world’s 8bn people live in a city. By the mid-century, that share will rise to three-quarters of the 10bn people expected to be alive. Urbanisation is accelerating fastest in South-East Asia and Africa, where entire cities are being built from scratch, and half of the world’s population lives. It means that every week a million more people live in a city than they did the week before.
This rapid urbanisation has a significant carbon cost. While cities generate roughly 80% of global GDP, they also create about 80% of greenhouse gases. They are increasingly vulnerable to climate extremes: from the urban heat islands of India to flash flooding in Japan and blackouts across Spain and Portugal. As a recent study by Professors Katerina Michaelides and Michael Singer notes, climate change is bifurcating city risks—drought in some regions, deluge in others, with an increasing amount of flip-flopping between the two. And so we face a paradox. AI might help optimise the city—but only if the city can power, hydrate and absorb the demands of AI.
Aligning intelligence with resilience
The solution is not to reject AI, but to govern it wisely and align its use with the goals of regenerative urbanism. AI can be used to enhance water recycling, optimise microgrids and forecast extreme weather. It might support better decision-making, rather than merely automate it. But this will require new governance, clear data rights, breakthroughs in quantum computing, and above all, sustainable energy sources—perhaps small modular nuclear reactors, or geothermal systems—to avoid AI cannibalising city infrastructure.
If we are to call this “AI for Good”, then “good” must be defined not by efficiency alone, but by equity, ecology, endurance and urban resilience. Cities have endured by adapting to shifting sources of power—from windmills and water wheels to steam, from fossil fuels to renewables. The question now is whether they can adapt to a new kind of power: artificial intelligence.
Author bio
Paul Wilson, chair, Smart Cities World
Paul is chair of the Advisory Board for Smart Cities World, read by 1.5m people worldwide, Smart Cities World provides daily news coverage and weekly features about the fast-changing world of urban innovation. Paul is also Senior Advisor at Open Agile Smart Cities with 170 city members and an honorary Research Fellow at the University of Bristol’s Smart Internet Lab. Until recently Paul was Chief Business Officer, Connected Places Catapult, the UK’s innovation accelerator for cities, transport, and place leadership.
Paul draws on 30 years’ international leadership experience in technology and cities. Earlier in his career he was chief marketing officer of $5Bn fintech company SunGard. He then led the development of the first economic growth strategy for the Bristol and Bath City Region of the UK, winning £260m of investment from the Government. He then founded the pioneering Smart City Project ‘Bristol Is Open’, winning numerous international Smart City awards and became a founder and Board member of the UK Government’s UK5G. Paul became telecoms industry association TM Forum’s Smart City International Advisor and then its Chief Marketing Officer, creating the ‘City as a Platform Manifesto’, attracting 300+ signatories from Governments, Companies and Cities worldwide.
Paul is a compelling conference speaker and writer. He works with a variety of cities, technology companies and associations. He combines an indepth understanding of digital technology with a strong understanding of public policy processes. He has lived in London, Hong Kong and Geneva, and worked extensively in America. Today he lives with his family in Bristol, England.
Views or opinions expressed are those of the author and any individuals cited, and do not necessarily reflect those of Economist Enterprise or any other member of The Economist Group.