1 Jun 2017
Intelligent algorithms may be the future of automation, but can they replicate human creativity? Hamish Nevile MEngNZ discusses whether machines will ever be a match for the spark of human creativity.
There are engineers around our Holmes offices experienced enough to remember a time before computers. At the other end of the scale, there are new graduates who can barely imagine a world without a screen and a keyboard. The advent of the computer age fundamentally changed the way the structural design world operates – and now we’re on the cusp of a similarly seismic evolution. Artificial Intelligence (AI) is here, and our world of work will be materially transformed by its arrival. It’s hard to accurately predict its impact industry-by-industry, but a study by the University of Oxford (2013) estimates that AI could replace nearly 50 per cent of jobs in the USA in the future.
I was lucky enough to visit Stanford University on an executive strategy and leadership course late last year. Much of the talk centred on how to adapt to the rapidly changing world of work, and how to deal with disruptive technology (think Kodak, Uber etc). No discussion on disruptive technology is complete without considering the rise of AI and the implications it could have for engineers and other professionals. The consensus was there are more unknowns than knowns, but the possibilities are boundless. AI will change the way engineers work and we need to proactively deal with its impact.
There’s an ever-increasing buzz around the development of driverless cars, un-piloted drones delivering pizzas and computers that can defeat chess Grand Masters. While those headline grabbing topics steal the limelight, there’s a groundswell of AI development that will do far more than delivering your meat lover’s classic in time for the start of the rugby (desirable as that is). In engineering, we increasingly use design automation to capitalise on the power of data and the good sense in automating repetitive tasks in the interests of expediting processes and eliminating human error. AI has the potential to go far further – rapidly learning as it goes to move on from repetition and automation and into intelligent design and problem solving. These are areas we have traditionally assumed needed ‘the human touch’.
So, will AI take our jobs?
It appears data driven roles—which could readily be replaced by intelligent algorithms—would be most ‘at risk’. A McKinsey research paper in 2016 estimated that 45 percent of activities in the US economy can be automated, and that’s only with the technology available now, not allowing for developments we haven’t yet conceived. To put it simply, in a very short time, anything that can be automated will be automated!
For example, software already rapidly iterates repetitive parts of the design process, rather than a person manually repeating it. At Holmes, we’re increasingly using techniques like parametric modelling with software like Grasshopper to rapidly iterate our designs. We’re able to manipulate data to interrogate different concepts very swiftly, cutting out a huge amount of repetitive work. It’s likely that AI could not only iterate the designs, but also find the optimal design pathways. Further, computer learning would gain experience from the process and use that knowledge to improve efficiency and outcomes.
It seems inevitable that engineering roles will, at the very least, change in focus. There will be more emphasis on smart programming and it will be more about the interrogation of data inputs and interpretation of data outputs. Where engineers of the past learned through accumulated experience gained on a huge number of projects, much of that learning might become irrelevant.
The realities of the coming wave of AI seem clear: our roles will change, even if they don’t disappear entirely. Perhaps I’m idealistic, but I just can’t imagine the wonderfully creative part of the design process performed by a machine. When we sat down with Gary and Nick at Stevens Lawson Architects to discuss ‘The Gateway’ project on Waiheke Island, the spark of human creativity and collaboration turned a bold idea into a ‘napkin drawing’, which we turned into a truly stunning structure. Similarly, the sensuous facade at the Len Lye Gallery in Taranaki was the product of dozens of creative meetings where vision, ideation and challenge transformed a blank page into a masterpiece. Could a computer ever do that? Could a computer ever conceive what the great engineering visionaries like Ove Arup, Peter Rice and Buckminster Fuller did?
There are important parts of engineering that seem to rely on the distinctly human qualities of artistry, creativity and collaboration. It’s the ability to dream, to think of possibilities and opportunities, rather than designing by programmed restraints or parameters. Not all engineers are created equal. What separates them is their problem solving, creativity and ability to move beyond the prescriptive nature of engineering conventions, parameters and standards.
Ove Arup himself valued and articulated the artistic demands of engineering: “Engineering is not a science. Science studies particular events to find general laws. Engineering design makes use of the laws to solve particular practical problems. In this it is more closely related to art or craft: as in art, its problems are under-defined, there are many solutions, good, bad or indifferent. The art is, by a synthesis of ends and means, to arrive at a good solution. This is a creative activity, involving imagination, intuition and deliberate choice.”
AI evangelists assert an intelligent computer will, one day, handle all of the process – artistic and scientific – and they might be right, but I’ll need to see it to believe it. Until then, humans will always be the single most important contributor to the design process, and I think AI can help us strike the right balance between the science and art of engineering.
At Holmes, it’s the creative problem-solving that’s always been our cornerstone, and what inspires us heading into a project. If AI and design automation deals with somewhere near 50% of repetitive engineering tasks, it could potentially free up 50% more time to focus on new ideas, new techniques and new possibilities. For us, that’s nothing short of exciting.
Hamish Nevile MEngNZ is the CEO designate at Homes Consulting. He has more than 20 years of experience in New Zealand and internationally. Following graduation from the University of Canterbury, Hamish joined Holmes Consulting. With Holmes and during his time working overseas, Hamish has delivered hundreds of challenging design projects all over the world. He has specialist skills in seismic design and his broad experience includes project work, competitions and detailed design and execution.