By Tom Cheesewright (@bookofthefuture), Futurist

12 March 2014: People are complicated. That’s why we haven’t built a convincing artificial intelligence system yet. Cities, as collections of millions of people, are therefore rather hard to understand.

But we’re optimists us humans. So we just keep trying. And we might be close to cracking it.

Knowledge is Power

The potential advantages of being able to truly understand a city are enormous. We are an increasingly urban species. But our urban environments face huge challenges. Climate change and its contributing factors. Commerce and the collapsed economy. Poverty. Transport. Energy. Education. The people running cities have to address every one of these issues. And the tools and resources at their disposal are limited.

This is why smart cities have caused so much excitement. One of the big things city bosses have lacked is information, particularly real-time information that allows them to make changes in a time frame of hours or days, not months or years. Sprinkling networks of sensors across a city can allow bosses to understand patterns in everything from traffic to temperature and respond appropriately.

There is no such thing outside the realms of fantasy. But it should give sufficient probabilistic resolution to be valuable. And that is a very good start.

Data, of course, is not even half the battle. Interpreting that data, and understanding the effects that changes you make may have on future behaviours of complex people, is where the real challenge lies.

In his book, Smart Cities, Anthony Townsend outlines the history of our attempts to tackle this challenge. From the early, simplistic attempts at modelling by Forrester, through to IBM’s more recent experiments in Portland. In Townsend’s eyes, none of these attempts has been an unqualified success.

Process and Presentation

The problem is one of both process and presentation. How do you create an algorithm or engine that can model the behaviours of millions of individuals, each with their billions of neurons, moods, desires and idiosyncrasies? Even if you can, how do you present this model back to city planners in such a way as to allow them to interpret it?

Ulysses Sengupta is willing to try. An architect and academic at Manchester’s School of Architecture, he has worked on modelling the rapid progress of India’s slums and China’s new cities, Sengupta is applying the multi-disciplinary practice of complexity theory to the problem. This complexity-based approach accepts that cities are not systems. They are not even systems of systems. They are a multi-layered, interactive maelstrom of thoughts, deeds, processes and plans.

Accepting this and modelling it are two different things. But a fourth-order complex adaptive system may be as close as we have come. The key word here is adaptive. Sengupta’s proposed model is not a static picture of a city, but one that interacts with its inhabitants, both via the data trail they leave behind unconsciously as they go about their daily lives, and consciously via social media. By capturing measured reality and considered opinion, Sengupta and his team hope to refine the model to the point where it becomes truly useful in decision making.

This will never be a precise view of the future: there is no such thing outside the realms of fantasy. But it should give sufficient probabilistic resolution to be valuable. And that is a very good start.