The Dark Green City: AI’s Unintended Consequences

What would cities look like if they were built from scratch, from the internet up? This is the question being asked by Sidewalk Labs, the urban innovation unit of Google’s parent company Alphabet Inc. What’s emerging from this thought experiment is a new approach to city-building, one that sees urban districts as platforms for testing and refining technologies that improve quality of life. Sidewalk Labs’ mission, it claims, is not to create a city of the future, but to create the future of cities.

This approach demonstrates how cities are on the cusp of a revolution in urban technology driven by sensors, ubiquitous connectivity, artificial intelligence (AI) and advances in digital fabrication and construction. Nowhere is this more apparent than in Toronto, where Sidewalk Labs recently partnered with Waterfront Toronto, a government agency, to create a new kind of neighbourhood. The 12-acre mixed-use district along Lake Ontario aims to be a global benchmark for how an advanced “smart city” can be built from scratch, quickly and effectively, using data-driven technology. If successful, the project would represent one of the most comprehensive models of how data related to just about everything — from traffic congestion to noise to air quality to trash bins — can be used to not only guide a city’s ongoing operations, but also teach it to improve itself, continuously and without human intervention. “Technology,” Sidewalk Labs insists, “can help create complete communities that are highly interactive and accessible to all, freeing residents from the constraints imposed by the heavy infrastructure and spatial hierarchies of the last century.”

While this may turn out to be true, it reflects a wider, commonly held assumption: that as technology improves it will always be for the betterment of humanity. Nicholas Carr, a US journalist who writes about the intersection of technology and culture, believes that this narrative of inevitable progress clouds our real relationship with technology. What we take as incremental improvements in our everyday lives may in fact obscure more nuanced and ambiguous changes.

What happens, he asks, when technology moves beyond lifting genuine constraints and starts freeing us from those things that we should not want to be rid of? Questions about data privacy and digital governance aside, Carr argues that liabilities associated with some technologies — in particular AI and automation — may become so advanced that they ultimately threaten to impair the conditions required for us to pursue meaningful work and meaningful lives.

As AI improves, it makes work faster and more efficient, and can lower environmental impact. It is estimated that a fully automated factory, without the need for lighting, heating or cooling, could operate using 35% less energy than a conventional one. But the products of its labour also require far fewer workers: that same energy-efficient factory could cut labour costs by 80%. So while the factory of the future may be green, it is also dark.

It’s not just physical jobs that AI is poised to take over. It’s also begun to encroach on a wide range of activities that demand intellectual judgment, from medicine to law. The rate of this intellectual outsourcing shows no signs of slowing down.

Historically, technology has delivered us with exponential gains in quality of life. The advent of something as novel and powerful as AI has the potential to deliver even more benefits. But it also has the potential to disrupt cities, and entire societies, with unintended consequences. Technology has also always been a double-edged sword; only now, the blade has become that much sharper.

Mark Bessoudo is a global research manager at WSP in London. He is also founder of platoforplumbers.com