AI, Cities and Climate Change: Whoever Said It Would Be Easy? - NewCities, Josef Konvitz

Submitted by Mike Osborne on June 16, 2019 - 14:19

Josef Konvitz, PASCAL Chair and Associate Fellow, Resilience, NewCities, attended a day-long conference on 21 May last month, sponsored by the NewCities Foundation, on the theme of Artificial Intelligence and Climate Change. Held in the sumptuous grand hall of the Paris City Hall, the subject highlighted the importance of public actors and local innovators. He wrote this report which appeared on the “New Perspectives” page of the NewCities Foundation on 14 June 2019.

AI, Cities and Climate Change: Whoever Said It Would Be Easy?

Josef Konvitz

Source: NewCities

The fate of cities, whether in Africa, Asia, America or Europe, is increasingly linked to climate change. Natural disasters – heat waves, floods, drought, earthquakes, violent storms – are more frequent and costly. Told that the goal of limiting the increase in CO2 emissions by 2050, and the rise in annual average temperatures, is unreachable, some people will be fatalistic, and resist changing their lifestyle. In several countries, far-right parties promise security behind borders in nations where power is centralized came first or second in the May 2019 elections to the European Parliament. But the Greens made impressive gains in others, often capturing three times as many younger voters than other parties – especially in cities. Politics matters as much as technology and science.

Into this picture comes artificial intelligence, AI. At a day-long forum on this topic convened in Paris on 21 May by NewCities – urbAIn: AI, Cities, and Climate Change, innovators presented new apps and systems, many on transport and energy; executives and politicians discussed the challenges of adoption including regulation, policy, and budgets. AI will give people choices, but they remain free to chose. Some will do more for the environment, others less – a classic free-rider problem when public goods – air, water, even public space – are involved. The market, after all, is democratic.

And here we come to the first of several barriers to the take-up of AI: people, or rather their individualism, their differences by age, class, income, culture, education, and their political views and values. A less polluting, carbon-neutral system is difficult to translate into everyday life when most of us face competing priorities at every step. Progress in cities may be cancelled out by trends in rural areas.

AI is actually increasing the demand for electricity. Infrastructure is a hot topic in political campaigns in many countries because many systems are obsolete, or are at capacity, or break down due to delayed maintenance. Yet the level of investment remains insufficient, by some estimates by as much as 30 percent, and this without taking into account what needs to be spent to cope with climate change. Nor is this all: in the fight for resources, peripheral and declining regions risk being short-changed. How much should be invested in them? And what is the cost of under-investment in growing, dynamic city-regions?

Immigrants are often entrepreneurial and innovative. Surveys tell us that people want more innovation but fewer migrants, a classic case of trying to have it both ways. Cities provide a hospitable environment for start-ups, some of which will bring new ideas to the market which will help solve urban problems. The transfer of innovation, and the skills to nurture entrepreneurship and to scale up promising innovations, call for fewer barriers to the trade of goods and services, and to the migration of people. Not exactly what we are seeing in the world today.

Regulation and regulatory policy was first mentioned at the NewCities Forum on AI by Thierry Mallet, CEO of Transdev, and it was also picked up by others. Based on experience, the evolution of regulation takes time, lagging behind the evolution of technology. Advanced governments are still digesting the regulatory reforms needed since the 1990s to cope with privatization, public-private partnerships, and globalization in the provision of public services. As we saw in the 1990s, problems emerged which led to public resistance and push-back. AI and the digital economy more broadly raise new questions about what is a good and what is a service, definitions which have huge implications for legal, regulatory and trade frameworks. The Organisation for Economic Co-operation and Development’s (OECD) Regulatory Policy Division has launched a major study into the ethical and practical challenges posed by AI for international regulatory co-operation. And on 22 May 2019, 42 countries adopted new OECD Principles on Artificial Intelligence to promote transparency, accountability and inclusive growth.

Even building design has a political or policy dimension. Many older buildings – stranded assets, in economic terms – are difficult to adapt to function with new environmental controls, domotics, etc. The rate of rebuilding in cities is in the range of one to two percentper year, a lot in a decade, but not when the goal is zero carbon by 2050. It is one thing to scrap cars more than 10 years old that have already depreciated; buildings costing millions however can have a useful life over centuries. Much better and easier to build in new systems literally from the ground up, in new construction. How adaptable will newly-built office, commercial and residential structures be in 20-25 years? Research and development is being invested in high-tech heating, power and water systems. Is enough being invested in the training of the next generation of craftsmen who will be called upon to maintain and upgrade these systems?

Faced with the prospect of an environmental disaster, the rational response is to do something to avoid it. People are rational but not in a uniform and collective way, especially against a target date of 2030 or 2050, too remote to inspire fear. When a disaster comes, most people will be unprepared. Will AI help cities be more resilient in the event of a catastrophe? There are two challenges. The first is for people – not machines – to take initiative when confronted by a dramatic event for which they are unprepared – think of the pilots of a Boeing 737MAX when the MCAS system put the aircraft into a dive. And second, disaster-prevention is based on memory. What people learn and retain matters more than machine learning.

AI has a great future but at the same time it forces us to examine and improve our institutions and policies, how the public is informed and consulted, and how experiments and innovations are evaluated for their impact. This calls for dialogue, for the kind of face-to-face discussions which NewCities supports and facilitates. The late Jean Gottmann, a geographer who worked in France, the United States and the United Kingdom and coined the modern term megalopolis, was a pioneer in the study of communications in cities. He was fond of saying that half the phone calls were to ask someone to meet for lunch. Now that Uber Eats and other services bring us our lunch, we need to remember how important it is to get out into the city, to see others at work and at play, to talk things over with friends, to do something un-planned, and perhaps, who knows, to return to work with a new idea, and maybe a new friend.

Josef Konvitz Konvitz joined the Urban Affairs Division of the Paris-based Organisation for Economic Co-operation and Development (OECD) in 1992 after nearly twenty years on the history faculty of Michigan State University. Upon retiring from the OECD in 2011 as Head of the Regulatory Policy Program, Konvitz was appointed Honorary Professor, University of Glasgow. Cities and Crisis, his fourth book, was published in January 2016 by Manchester University Press (distributed in North America by Oxford University Press). While at the OECD he analysed the territorial impacts of large-scale disasters, including floods in France and earthquakes in Turkey and Japan.