This relative urban tranquility could be challenged by the thundering arrival of AI. The term cognitive city is already on many people’s lips, and it is driven by AI agents that can communicate with each other to anticipate residents’ needs (multi-agent AI). But as is often the case with brand-new concepts, it is important to distinguish the promises from reality.
Cities that communicate
While the fantasy of a responsive, autonomous city turbo-charged with tech is still quite far from reality, effective, one-off generative AI use cases are on the rise. The most common being the use of intelligent agents, which simplify access to urban services, thanks to improved information retrieval. This communicative city holds the promise of “de-bureaucratization” for residents.
For example, in 2019, Buenos Aires pioneered the way forward by launching Boti, a chatbot – or conversational agent – that achieved a monthly average of five million conversations on topics as vast as mobility, recycling, health, security, culture and tourism. In France, Issy GPT offers a similar service and does not hide its affiliation with major language models such as GPT. The city of Singapore went a step further when it launched AI Trailblazers, an initiative aimed at identifying one hundred generative AI use cases for the city within one hundred days. The result? Forty-six of them have already built a Minimum Viable Product (MVP).
These include a chatbot that can retrieve economic data from the National Economic Research and Visualisation Engine (NERVE). Another is a partially automated course content creation system for academic staff. And a final, more simple project is one which aims to make it easier to book a badminton court by indicating availability.
Services and governance over economy and business
These examples illustrate a fairly obvious trend: a large majority of today’s urban AI projects aim to improve urban services and governance. A study conducted worldwide by the Barcelona Centre for International Affairs (CIDOB) shows that 66% of AI projects concern the provision of services and governance. Trailing behind this are mobility, social services and resource management, which concern respectively 24%, 22% and 22% of projects. While urban planning is often presented as an area benefiting most from AI, here it only reached 11%, while the economic sector is barely represented at all, with 4% of initiatives.
Image Credit : Alexsl Getty Images Signature
Is transformative AI the future?
The emergence of a truly transformative AI, capable of proposing new urban practices, remains relevant despite all this. Some projects are trying to go beyond the logic of pure optimization and propose the odd futuristic solution. For example, artificial intelligence is a sine qua non condition for the development of autonomous drone delivery services, which raise questions concerning adapting to a uniquely complex urban environment.
Developed in Coventry in the UK, Air-One is something of a pilot project for the topic. Outdoor advertising in cities could also be completely transformed by offering personalized content based on the viewer’s characteristics or campaigns adapted to the real-time conditions (such as the weather, for example). The climate crisis is also finding promising solutions in AI. For example, Google is using image analysis to help understand heat islands, to identify the best places to plant trees, and to optimize traffic light timing configurations to reduce emissions.
Fear over the city?
A certain number of barriers are yet to be broken down if AI’s promises are to become a reality. For example, 74% of mayors cite barriers such as insufficient technical expertise and 70% cite budgetary constraints. Meanwhile, 62% highlight data privacy and security concerns, and 47% cite ethical concerns. These last two issues are embodied in disputed practices, such as the use of facial recognition by the police in London, or the development of predictive policing, to anticipate “hot spots” that are high-risk areas for crime. Using AI to monitor certain critical infrastructures also raises security concerns.
The World Economic Forum stresses the importance of establishing governance frameworks for AI, in what still seems like an “exploratory phase.” It gives the examples of the algorithm register established in Amsterdam, which offers an overview of all the AI systems used by the city, or AI user guidelines created by the city of Seattle. While there is currently an abundance of enthusiasm, a lot still remains to be done before AI finds its rightful place within our cities.
