Are smart cities all that clever?

The governance of data is another significant public concern. People are rightfully asking who will access, store and use the data gathered about them. While keen on connectivity, people simply aren’t so willing to give up their privacy.

In 2017, Dan Doctoroff wondered, “What would a city look like if you started from scratch in the internet era – if you built a city ‘from the internet up’?” At the time, Doctoroff was head of Sidewalk Labs, an ‘urban innovation’ division of Alphabet (then Google). The company was becoming involved in developing the Quayside area of Toronto’s waterfront.

In her book, she argues that ‘smart’ computational models of urbanism advance an impoverished understanding of what we know about a city, and of what’s worth knowing. Instead, she argues, we should think beyond smartness to the other types of intelligence inherent in a city, such as the knowledge embodied in a library or the experience gained through care work.

Mattern is fond of the term ‘grafting’ as a more considered way to apply new technologies to an existing system. “If we think about applying new technologies through this artisanal craft-oriented method of suturing, we can imagine a more thoughtful way of integrating new technologies into an urban environment where people are already living, where there are forms of intelligence circulating among those populations, and where existing technologies are already working quite well for them.”

Overall, rather than seeing a city merely as a computer, she encourages us to view its many facets. We should open our eyes to the multitude of metaphors it embodies, whether a tree, a computer, or a graft, when considering how to develop more robust approaches to urban development in future.

Data concerns are one of the major criticisms directed at smart cities. In addition, there has been considerable criticism of the racial and gender biases that can be inherent in technology, such as surveillance cameras or criminal justice algorithms.

For example, an algorithm widely used in US hospitals to allocate health care to patients had been systematically discriminating against black people, according to an article in Science (October 2019). Similarly, predictive policing algorithms have been accused of racism. Artificial intelligence has proved to be gender-biased in instances, including setting credit limits significantly lower for women on similar incomes and expenses to their husbands for no explainable reason.

It’s especially important to be careful when handling the data of historically vulnerable communities, says Mattern. “When dealing with those who have been historically mistreated through the collection of data, special attention has to be paid to equity – to the particular needs and desires and values of those communities.”