Daniel Susskind: ‘Automation of jobs is one of the greatest questions of our time’

Source: The Guardian – Artificial intelligence

Daniel Susskind is an economist and fellow of Balliol College, Oxford. He has held policy roles in the Blair and Cameron governments. His new book, A World Without Work, explores how society should respond to the increasing automation of employment.

This isn’t an unexplored topic, so why did you write this book?
My view is that this is one of the greatest questions of our time. And in spite of everything that has been written, I didn’t feel like we had done the question justice. I don’t think we’re taking seriously this idea that there might not be enough well-paid work for everyone to do because of technological advances that are taking place.

Is that partly because of the way the question is framed, that “robots are stealing our jobs” when that’s not how these changes are happening?
Yes. There are two mistakes there. One is the sort of anthropomorphisation of technology – these robots exist but they are often gimmicks. The technologies that are really very powerful don’t look, think or reason like us. The second mistake is to think of entire jobs being replaced. These technologies tend to displace people from tasks – what I call task encroachment.

Many of the boundaries economists and computer scientists developed in the second part of the 20th century for thinking about what machines could and couldn’t do have been crossed. For example, driving a car, making a medical diagnosis or identifying a bird from a fleeting glimpse. All these tasks can be accomplished by software now.

And machines are often performing these tasks using strategies that are new to humans?
We thought many of these tasks would be difficult to automate because humans couldn’t articulate how they performed them. They relied on experience, intuition and gut reaction – so how could you write a set of instructions for a computer to follow? But by using lots of data and computing power, machines are creating new strategies.

In the light of this technological change, what career path would you advise a 16-year-old to follow?
Very crudely, I’d say, there are two strategies: either you learn to be good at the sorts of things these systems and machines cannot do or you try to build the machines.

But forecasting which jobs and tasks will be automated is hard. Should people bother training to be doctors or lawyers when automation is already encroaching on these careers?
For young professionals, perhaps somewhat counterintuitively, I say the best thing to do is pursue the traditional path. But be far more agnostic and open minded along the way about opportunities that come up. Because what’s interesting is that if you look at these technologies – take the systems developed by DeepMind or the ones that recognise melanomas by Sebastian Thrun’s team at Stanford – these teams contain lots of domain experts, such as trained doctors.

Technological unemployment appears be growing alongside rising inequality – is there a direct relationship between the two?
Work is our traditional way of distributing income, so one of the great challenges of a world with less work is how we slice up the pie. The labour share of income is falling in many parts of the developed world.

Should robots be taxed and that money funnelled back to the victims of automation?
There are various problems with this idea. What do we mean by a robot? A driverless car isn’t driven by an android. How do we do a robot headcount? We do need to tax capital, the owners – however that has a branding problem.

But the owners of the these technologies are often quite good at avoiding tax.
If the trends in inequality continue, focusing on how we resolve that problem is going to become more pertinent.

You advance the idea of a “conditional” universal basic income.
I think the challenge in a world with less work is how you maintain that sense of social solidarity. At the moment, that comes from a sense that everyone is paying into the collective economic pot through their taxes or, if they’re not in work, they are actively looking for work or training for work.

I think one of the problems of a universal basic income [a payment made to the entire population regardless of employment status] is that too many people take offence at the idea that you give something and don’t expect anything in return.

But perhaps we can introduce conditions to a UBI, for example that you undertake some valuable voluntary work in return. There are something like 15 million people doing a hugely valuable set of voluntary activities in the UK – why not recognise that? The UBI solves the distribution problem – how do you share income if the labour market doesn’t do it very well? – and if everyone is contributing, some in economic and others in non-economic ways, that may solve the we’re-all-in-it-together problem.

Volunteering requires empathy and hands-on skills that computers find hard and such activities are not rewarded very well.
There is a paradox. Machines can’t do them and they tend to be unpaid by the market. There’s this huge gap between the social value of a lot of this work and the value that’s recognised in the market. So one of the reasons for optimism is that there’s an opportunity to potentially address that.

Work not only gives people income, but often gives meaning and identity. How can people get meaning from more leisure?
If we think the relationship between work and meaning is very tight, the alternative might be not to think of this as being about the future of work, but the future of leisure. You know, we have labour-market policies to shape how people live their working lives. Maybe we also need leisure policies to shape how people spend their spare time. We already have a pension system that is a heavily subsidised leisure policy of sorts.

You’ve worked in government. Is automation a priority for politicians or are they reluctant to address things that may only pay off long after they have been voted out?
Almost every government in the developed world has published some kind of AI strategy in the last few years. Whether or not they’re then following that strategy is less clear.

You have to explain that this is not a story about the future – this is happening now. Every day, we hear stories of these technologies, driving cars, making medical diagnoses, drafting legal documents, designing buildings – this is something that is happening now; it’s not in the future. It is closely linked to the inequalities we see emerging. It tends not to be acted upon because it’s thought to be something that doesn’t really matter, but it does.

A World Without Work by Daniel Susskind is published by Allen Lane (£20). To order a copy go to guardianbookshop.com. Free UK p&p over £15