UdeM faculty members helped write a white paper on how AI systems could support the development of socially and ecologically sustainable cities.
A white paper on the risks and benefits of using artificial intelligence in urban development has been published jointly by Mila, the Quebec Artificial Intelligence Institute, and UN-Habitat, the United Nations program to promote sustainable urbanization.
The paper is about how to ensure that the deployment of AI-based technologies in urban settings will reduce risks to individuals and the environment while maximizing potential benefits for their populations.
It contains findings and recommendations for municipalities, governments, urban planners and technical professionals.
The project was led by several members of Mila and Université de Montréal, including Shin Koseki, a professor in the School of Urban Planning and Landscape Architecture, Jean-Louis Denis, a professor in the School of Public Health, and Catherine Régis, a professor in the Faculty of Law.
Applications in secen sctors
The document focuses on urban applications of AI in seven sectors: energy, mobility, public security, water and waste management, health care, urban planning and urban governance.
It describes how AI could:
- accelerate the transition to a low-carbon society by predicting energy supply from variable sources and amounts generated by wind and weather;
- forecast demand for ride-sharing systems;
- advance the adoption of electric vehicles by predicting battery condition, degradation, and remaining life;
- analyze meteorological data to predict flooding and inform home buyers and people purchasing insurance;w
- optimize vehicular and pedestrian traffic;
- detect toxic contaminants to improve waste management systems;
- and create remote health monitoring systems for measuring heart rate, body temperature, blood pressure, etc. to support real-time interventions.
‘Significant human-rights issues’
“Artificial Intelligence is a tool with enormous potential for fostering sustainable urban development, but it also raises significant human-rights issues,” explained Koseki, the UNESCO Chair in Urban Landscape at UdeM. “The white paper provides guidance for assessing where and for whom AI applications add people-centered value and are appropriate.”
The authors of the white paper have left no stone unturned in their analysis of the risks associated with each potential use of AI. The disadvantages take many forms: inequity, bias, substitution of automated systems for human beings, breach of confidentiality, misalignment of data with reality, political repression, and more.
“There are risks at every stage of the technology’s deployment but particularly when selecting the data for running algorithms,” said Koseki. “To avoid bias, I would stress the importance of having staff who are knowledgeable about and sensitive to diversity issues, who themselves have a range of technical knowledge, backgrounds and experience.”
The Mila/UN-Habitat white paper AI & Cities: Risks, Applications and Governance is free to access and download online.