Will robots take over finance roles?

Robot office workers assume finance rolesCould Optimus Prime or C-3PO take over your job on the bond trading desk? Perhaps not tomorrow or even next year, but one day in the not too distant future robots could be overseeing the workings of the financial markets in ways we can barely fathom today.

Already algorithmic trading is beating human beings to the punch, as computers can execute trades in nanoseconds. So why not take this further and have robots take over more of the decision-making functions? Why do we need humans to manage the algorithms? 

Finance jobs

To a great extent, the argument about robots wiping out the labour market depends on the type of role being performed.

Jobs can be split into those that are cognitive and manual, and routine and non-routine. Like a lot of white-collar industries, finance includes a lot of the cognitive and routine roles for which robots are ideally suited. On the other hand, manual roles are physically hard for robots to perform as automation technology is just not that advanced, while non-routine cognitive roles remain too complex.

Repetitive cognitive roles, such as those in middle management or administration, are arguably most at risk. This produces an hourglass effect on the jobs market, with those at the top and those at the bottom unaffected while the middle gets squeezed.

“People who use their hands will have jobs for life,” Dave Coplin, a senior Microsoft executive, said in an interview with The Telegraph. On the other hand, bankers, traders and many more in the financial sector could find it tougher.

Moravec’s Paradox helps explain the position. “This states that what we think is easy, robots find really hard, and what we think it really hard, robots find easy,” explained Coplin.

“Complex maths equations are hard for humans but take nanoseconds for a computer, but moving around and picking things up is easy for us, while being almost impossible for a robot.”

Should we worry?

Concerns about machines taking jobs stretches back centuries. The Luddites smashed the power looms that did them out of a job in the early 19th century. They found new work, so what’s all the fuss about today?

As some argue, progresses in robotics and computing mean humans face a unique challenge in the 21st century - revolution not evolution.

In a new paper in the Journal of Economic Perspectives, Gill Pratt suggests technological developments are fomenting an "explosion in the diversification and applicability of robotics". 

He argues: "Many of the base hardware technologies on which robots depend - particularly computing, data storage, and communications - have been improving at exponential growth rates."

Advances in 'cloud robotics' (learning from each other) and 'deep learning' (processing huge amounts of data to improve their capabilities) mean the machines are getting much more adept at performing a wider range of functions.

Improved automation makes robots better at manual tasks, while artificial intelligence advances mean machines can take on ever-more complex cognitive tasks. As Mr Coplin points out, “the City could be run by algorithms”.  

University of Oxford researchers Carl Benedikt Frey and Michael Osborne reckon 47 per cent of total US jobs could be taken over by computers by 2033. Meanwhile, more recent research by the university estimated that up to 35 per cent of British jobs could be replaced by advances in computing and robotics over the next 20 years.

But those that worry about the rise of the machines ignore simple truths about labour market economics over the last hundred years. David Autor from MIT argues that technology also creates tonnes of new jobs, at least as many as it replaces. For example, he notes how the number of bank clerks in America rose between 1980 and 2010, in spite of ATMs and other banking technology spreading.

Robots may not be taking over just yet, but the skills humans will need in the future will undoubtedly have to modify as automation and computers do more.