Thinking robot

+ Designers and engineers need to shape how artificial intelligence will help humanity make better-informed business decisions in the future.

Does your work involve making strategic choices? Now is your chance to shape how artificial intelligence will help you make better-informed business decisions in the future.

Computers are already helping make hundreds of decisions every day. Designers and engineers use optimisation techniques and parametric modelling to shape our designs. Like me, you probably rely on the internet and your smartphone to find the quickest route through a city. Natural language recognition, image recognition and other futuristic tools are now available to consumers. And just recently, a machine passed the Turing test for artificial intelligence – the first win in 60 years.

So can we use a machine to help us in making strategic choices too? Can a machine be used as a digital assistant who estimates and scores a series of decisions on the basis of the possible consequences and outcomes? Can a machine help us with forecasting and implementing a better future?

I think it can and it will; and it’s time that designers and engineers engaged with this. Put simply, if you want help making a decision, you just need to feed the right data into a machine that has the ability to learn. 

You’d start with a model of reality that is described by parameters, whether that’s elements of a building, people in an organisation or anything else. Think of buildings and the evidence contained in a design. We can measure several parameters from an existing building – things like temperature, irradiance and running costs. We can look into the process that produced the building – the design meetings, the sequence of decisions.

We could then feed this data into a machine and start asking questions. Is a large glazed area a good idea? What are the design consequences of this choice and what is the likely cost of the building? And how would it affect acoustic performance?

The machine would have an optimisation layer so you can score different options, and an algorithm to refine the design until you find the best outcome. Because decisions are never simply true or false, the machine would also need a decision-making system based on probability. And finally, it would need a dataset from previous situations to help it make decisions.

The power of this tool is anticipation. We would not use a machine because it is fast; we would use it because it can allow us to predict, to make choices about the future with a better understanding of the consequences. Of course, we would need specialists to prepare the right data and fine tune the machine’s logic. So engineering would need to evolve.

That’s why I think it’s time for a wider audience to get involved in understanding and influencing this emerging area of computing.  As designers and engineers, we need to be getting our hands dirty, digging into code and starting to write our own tools for decision-making.