#WorkingWorld
'Forget brexit and immigration, the robots are going to take all our jobs!' There is even a website to tell you the chance of it happening. Or, if you prefer more measured hype, 'robot automation will take 800 million jobs by 2030'. Daugherty and Wilson, writing in 'Human + Machine', are critical of the Man vs. Robot mindset that underpins such headlines. It makes great scifi, but misleads when we ask what to do about the rise of AI in the workplace.
How might work change?
So is AI eating the working world(1)? The report the BBC cited spends much effort on how work will change, and concludes that 'a very large number of people may need to shift occupational categories and learn new skills in the years ahead'. Daugherty and Wilson talk persuasively of 'reimagining process'. Rather than training for a different but existing role, the real challenge is to reimagine the roles themselves, to reimagine the way we do things: with technology not a dumb tool, but a smart colleague. That seems to ring true for the professions, if only they are humble enough to recognise that skills we view as profoundly human are susceptible to improvement and augmentation with AI. Radiology is a nice example. Deep learning AI systems are able to match or exceed human success rates when assessing medical images, and the technology will not stand still. At the process level, the challenge is to reimagine how to look after patients given these new capabilities, and the upside of ever-improving healthcare is obvious and desirable. At the individual level though, what happens to radiologists?
The importance of explaining AI
Entirely new roles are emerging with the rise of AI and deep learning in particular. One fascinating grand challenge is to explain how AI makes decisions. There is inherent opacity in sophisticated algorithms, which throws up a very human reaction. IBM's Watson platform is used in evidence-based medical diagnosis and when its predictions agree with the experts, all is fine. When Watson differs though, the human instinct is to say it is 'wrong', even where Watson has found correlations in the data unrecognised in conventional practice. Explaining how AI arrives at its answers will be critical to advanced adoption.
What will society look like?
With all the reimagined will in the world though, it seems to me that Keynes had a point. In 1930, he predicted that workers in developed countries would need work only 15 hours a week. The challenge would be how to use one's leisure time wisely. Many jobs will disappear to automation, and the displacement of workers might be too great to be offset by any amount of retraining. Or better put, if we no longer need deploy so much human effort to have a good society, can we (or our politicians) imagine a better way of doing things? It seems that radical society wide changes, such as a universal basic income, and planning parts of the economy must rise to the top of the agenda. But as Keynes said, 'the difficulty lies not so much in developing new ideas as in escaping from old ones'.
If you would like to know more about this topic, please contact our Innovation Manager: Philip Lindan.
