Daron Acemoglu on a robotic future
| Datum: | 18 januari 2017 |
What will happen to jobs as more tasks are done by robots? Will mass unemployment ensue, or will humanity adjust as it has to new technologies in the past? The answer is uncertain, according to leading economist Daron Acemoglu. The MIT professor laid out possible scenarios for the future when he visited FEB in November to give the Groningen Growth and Development Centre’s Maddison Lecture.
The title of his talk was ‘The Race of Man Against Machine’, and it touched on a key topic of research at the Faculty of Economics and Business: inequality. Inequality is the subject of one of FEB’s signature areas, which bring together researchers with different specialties to solve problems of international importance. Acemoglu’s lecture at FEB was particularly relevant. Here are key sections of the talk in Acemoglu’s own words, condensed for brevity:
Technologies can be ‘enabling’ or ‘replacing’
'Enabling technologies complement or increase the productivity of certain types of skills. An obvious example is computer assisted design. Yet increasingly we are in a new phase, where technologies don’t augment or enable workers and their activities, they replace them. Robots are the ultimate replacing technology. They are not helping people do tasks, they just take over tasks.'
Where is the effect?
'Our data is from 722 commuting zones in the United States, which is behind Europe in terms of penetration of robots. Exposure to robots is essentially entirely on the east coast of the United States, because that’s where a lot of the industries that have been automated or robotised are located. You can recognise some of these places as the heartland of Trump. The main industry that is affected by robots is automobiles, where assembly lines are filled increasingly by robots. But other industries are affected, including metal products, plastic products, chemicals, electronics, and food.'
What is the result?
'I get fairly large numbers: one additional robot per 1,000 workers reducing employment by seven jobs. Wage effects are also pretty large: one robot per 1,000 workers has reduced wages by 1.6 percent. You should be healthily sceptical about these results because it’s comparing trends in different cities. But the effects are there for men and women, slightly larger for men. You see the effects where you would expect them, in the heavily robotised industries.'
What future scenarios are there?
'Differently to other technologies, which were replacing some tasks but also augmenting other workers, robots are not directly helping anybody. Employers benefit, productivity increases, but labour does not benefit. So there are huge distributional consequences.
The optimistic scenario is that there will be employment creation but that it will come from new tasks. If robots are replacing a lot of tasks then there’s a lot of labour that’s being freed. Relatively cheap labour is going to fuel the creation of more tasks. From the 1980s to the 2000s, new employment often came from tasks which did not previously exist in the dictionary of occupations. Things like management consultants, radiologists, programmers, and app writers. The pessimistic scenario is that we are going to have more and more workers pushed into low-wage occupations. I think the future is very uncertain.
Perhaps we are going to create lots of unemployment, lots of low wages, but also a lot of wealth because of these productivities. But that’s uncertain, and there are also some puzzles. First, we have not so far found good ways of redistributing that wealth. Second, we don’t actually seem to be creating a lot of it; or not a lot of it in most of the industries where we expect it.'
Are there some tasks which robots will never do?
'We have learned not to underestimate computers. They are an increasing number of tasks that 10 years ago were thought to be impossible for machines to perform, and we now take for granted: beating Go masters, reading emotions, passing the Turing Test. Even five years ago I read many philosophers arguing that this was going to be impossible, and this is changing entirely.'
Daron Acemoglu, 49, is the author of Why Nations Fail, along with James Robinson. He was born in Turkey and studied at the University of York and LSE. Acemoglu is the Elizabeth and James Killian Professor of Economics at MIT and the fifth most-cited economist in the world, according to IDEAS. He is a recipient of John Bates Clark Medal for economists under forty judged to have made the most significant contribution to thought and knowledge.