Artificial intelligence (AI) is much maligned for its resource requirements. It consumes vast quantities of energy and water, in a worryingly increasing trend.
But scientists at the University of Helsinki are developing AI processes to slow climate change by informing smarter city design. To improve quality of life while reducing emissions, sustainability science professor Laura Ruotsalainen and her research group are creating machine-learning methods to simulate traffic flow, among other municipal features.
Sim City
Simulations are often used by engineers to explore civic solutions without relying on real-world evidence that may take years to gather. Therefore, urban planners can recreate different scenarios on exponentially quicker timelines, before a single brick is laid.
Accordingly, Helsinki researchers are using AI to improve traffic infrastructure to streamline smooth commutes, reduce emissions, and improve air quality.
But how do they do it, at a fundamental level? Through reinforcement learning. For perspective, imagine an AI entity driving around a virtual city. It tries many different routes then receives digital reinforcement when it achieves set goals — and penalties when it doesn’t.
AI agents can now do this at ever-greater scales and speeds to assess city-wide efficiency: “An urban environment involves a vast number of interacting effects. AI can help address such complex problems and the combined effects of multiple factors. This is work that humans cannot perform unaided.”
Optimizing Vital Civic Services
AI also helps create safer structures by allowing engineers to explore the impact of natural disasters. By modeling digital twins, or virtual versions of urban landscapes, researchers can examine the impact of floods and other catastrophes to inform better response tactics and save lives.
Other services are just as vital: “San Francisco, for example, has implemented a smart waste management system that uses sensors and internet-connected devices to optimize the collection and disposal of waste.”
Yet multiple challenges remain. Researchers must quantify, gather, and integrate vast amounts of data among varied, possibly incompatible formats.
Addressing these challenges will be worth the effort. Around 60% of the world’s population currently lives in cities, and that figure is forecast to grow to 70% by 2050. Accordingly, any technological tool that could help improve civic satisfaction and slow climate change is invaluable to both the population and the planet.