Computers and populism: artificial intelligence, jobs, and politics in the near term | Frank Levy, 2018, Oxford Review of Economic Policy 34:3, 393-417

Claims:

  1. AI will automate and eliminate blue-collar and clerical jobs
  2. In 5-7 years, we will start seeing the effects of this movement
  3. A new populist politics could be based on pitting "the people" whose jobs are automated against "the elite" such as the software developers that helped that automation.

Structure:

  1. How does AI replace, modify or creates jobs? Emphasis on machine learning
  2. Four example of current jobs where AI causes displacement
  3. What are AI applications that will replace jobs?
  4. How might this contribute to a new surge of populist politics?

How does AI change human work?

A job can be considered to be a set of tasks. AI do not replace "jobs" directly; rather, different tasks are automated.

Theory of task automation makes two observations (but I wouldn't say they are neutral observations):

  • all human work involves processing information
  • a computer processes information by executing instructions

I'm curious whether I could find an example where, in the broadest sense possible, work does not imply handling information.

Software uses two kinds of instructions:

1) Deductive instructions (rules) 2) Data-driven instructions

The latter is used when we cannot directly express the information processing structure.

An important difference is that option 1 often allows an error-free modeling of some information domain by expressing its rules and how to deduct knowledge. In the second case, you make predictions that may be wrong.

Polanyi's Paradox:

A perspective on this development comes from the writing of scientist and philosopher Michael Polanyi. A half-century ago, Polanyi wrote: ‘[W]e know more than we can tell’ (Polanyi, 1967), an idea that became known as Polanyi’s paradox. As an example, we can know how to ride a bicycle but we can’t explain to a child how to ride a bicycle in a way that keeps her from falling as she learns. By discovering information-processing instructions that we cannot articulate, machine learning allows us to unravel at least a part of Polanyi’s paradox. (398)

Cf. Wikipedia: "Polanyi's paradox has been widely considered a major obstacle in the fields of AI and automation, since the absence of consciously accessible knowledge creates tremendous difficulty in programming"

Nevertheless, machine learning still has clear limitations. For example, predictions are of low quality when a situation is encountered that was not covered by training examples.

How to resolve this?

  • Give the best answer available (which is likely very bad) and do not notify anyone about this.
  • Reinforcement learning: keep learning e.g. through user feedback.

At this point, however, the best guess is that AI is likely to make most near-term progress in automating narrow, structured tasks—structured because machine learning is predicated on identifying an underlying pattern of information processing; narrow because it is desirable to estimate the model on a training sample that contains most of the situations the software is likely to confront. (399)

Which jobs are vulnerable?

Summary: occupational polarization.

Taken together, these occupational descriptions help to explain automation’s role in the gradual polarization of the occupational distribution (Levy and Murnane, 2004; Buyst et al., 2018, this issue). Automation of the lowest wage jobs is hard because these jobs require unstructured physical activity and unstructured social interaction. Automation of the highest wage jobs is hard because the jobs (including many of the jobs AI creates) require unstructured cognitive activity and unstructured social interaction. Automation’s impact is relatively larger in mid-wage, mid-skilled occupations because they involve relatively higher levels of structured physical and/or cognitive tasks. These are relative distinctions. We have seen that part of a lawyer’s job can be automated and camera-enabled robots can automate part of the job of a security guard. But on balance, near-term AI will have the greatest effect on blue-collar work, clerical work, and other mid-skilled occupations (402)

So what is expected is a push of persons formerly doing those middle-skilled jobs, into lower-wage work.

How fast will AI advance?

What applications can we expect to replace some jobs that exist now?

  • Autonomous long-distance trucks
    • Feasible in quite regular environments, such as ports.
    • Truck platooning
  • Automated customer service responses
  • Industrial Robotics

Overall the estimation of the author is that the amount of job losses due to AI is very moderate compared to job losses due to Chinese import, general automation, recession. (cf. 410).

However, he also emphasizes that job losses due to AI will be covered in the media extensively. And when for example truck platooning is implemented, people will see those platoons on the road. This increased visibility, combined with the fear of other people that their job might also be at risk, set's the foundation for the next section.

Automation and populism

Pretext:

There is good evidence that manufacturing job losses, driven in part by Chinese imports, played a role in the 2016 presidential elections (e.g. Autor et al., 2017). Since automation anxiety is already high (Smith and Anderson, 2017), it is likely that AI-induced job losses will eventually create their own political reaction (410)

To me this pretext is not that strong.

Builds forth on columns by Jed Kolko.

First of course, what does the author mean by populism exactly? It is quite clear the author is targeting Trump, specifically.

"Populism describes a particular kind of group conflict by adding the ideas of victimization and illegitimacy." (410)

From Oliver, J. E., and Rahn, W. M. (2016), ‘Rise of the Trumpenvolk: Populism in the 2016 Election’, ANNALS, AAPSS, 667, September, 189–206:

Its goal is restorative, replacing the existing corruption with a political order that puts the people back in their proper place and that is more faithful to their longings and aspirations. Its worldview is apprehensive, suspicious of any claims to economic, political, or cultural privilege; for populists, the good is found in the common wisdom of the people rather than the pretensions of the expert.

Also:

Hannah Arendt on propaganda in time of economic and cultural dislocation, Arendt, H. (1976), The Origins of Totalitarianism, revised edn, Houghton Mifflin Harcourt:

What convinces the masses are not facts, and not even invented facts, but only the consistency of the system of which they are presumably part. . . . They are predisposed to all ideologies because they explain facts as mere examples of laws and eliminate coincidences by inventing an all-embracing omnipotence which is supposed to be at the root of every accident. . . . Totalitarian propaganda can outrageously insult common sense only where common sense has lost its validity. (Arendt, 1976, revised edn, pp. 351–2)

Why automation and populism may not be related

See 411-412:

  • Populist movement require specific conditions and are usually short-lived. Once AI-caused job losses are a big issue, the political climate may have changed.
  • Job losses due to AI are likely slower than for example a whole factory closing at once, and geographically distributed, and harder to politicize.
  • The aforementioned applications are in areas with big turnover, so these innovations might not lead to big layoffs.
  • Due to the usual ideology and rhetoric behind populism, they are more likely to focus on job loss due to cheap foreign labour / trade.

I'm actually quite disappointed by this paper.

This is more or less the only point:

These data suggest that a populist politician who campaigned on AI-induced job loss would start with ready-made definitions of the ‘people’ and the ‘elite’ based on national fault lines that were sharpened in the 2016 presidential election. This politician could also point to an evident lack of respect: a set of highly educated coastal ‘elites’ who make a very good living developing robots to put ‘the people’ out of work. (414)

Discussion points