Some of the course literature hints toward future developments. For example, Johnson (1997) asks if wage inequality is likely to continue rising in the future. He argues that inequality may be reversed when computers become self-sufficient, making skilled labour superfluous and increasing demand for unskilled labour. Goos (2018) is not that rigorous, but he also advises higher investment in occupations in non-routine social, motivational, and interaction skills, since they will probably be hard to automate in the near future. Autor (2015) thinks it will not get that far, mostly because of environmental control and the limits of machine learning, captured in the Polanyi´s paradox.
Personally, I also think fears of extreme job replacement may be a bit farfetched. Autor (2015) has a fair point with the Polanyi paradox, we do know much more than we know. However, I think the role of workers may change radically in the very near future. As several authors in the course literature said, IT does not always take over entire jobs, but is often able to take over specific, simplified tasks. The role of a worker may well evolve to some sort of an IT operator more and more over time.
Think of the 5 types of tasks of that Levy and Murnane (2013) develop, of which we already agreed that routine tasks are at risk of being substituted by IT. Manual non-routine tasks are hard to codify entirely, but specific tasks within jobs are easily codifiable. For example, think of health care for elderly, where many tasks are very physically demanding (therapy, washing, et cetera). Machines may take over the heavy tasks, and an ‘operator’ ensures that all goes well, or talks the elder person through the process. Lately, there have been tests in the Netherlands where robots were used to give some personal attention and assistance to lonely elderly, so even the human, or social part can be substituted for.
Meanwhile, some threatening developments are also going on for working with new information and solving unstructured problems. Artificial intelligence is characterised by its ability to adopt information and learn to see patterns. Pattern recognition is thus very much at risk of being substituted by artificial intelligence in the very near future. Because of artificial intelligence’s ability to learn exponentially, I expect that a certain degree of common sense will be achieved anytime soon as well. These developments enable IT to penetrate the high-wage occupations as well. Managerial optimisation decision are made quicker and more efficiently by computers, perhaps a manager is only needed to check the feasibility of the proposed solution. But then again, if this machine can learn inductively, and creates new rules for itself after being corrected, this manager is no longer needed in the future either.
Concluding, I think that routine-based technological change may seem like a good theory for the labour market present-day, but it will not hold for upcoming decades. Artificial intelligence and machine learning will grow exponentially and thereby change all existing models into something not imaginable yet. This may sound somewhat scary, but we are creating IT that is able to think autonomously, and that might have grave consequences for society as we know it. For the near future, maybe we will be much more reliant on IT, and some IT will be harder to operate than other technologies, so wage differentials may still prevail. But where we are in 30 years? That may depend on IT development, but also on societal developments. Think of the problems mentioned in blog II, but also about recent increases in part-time contracts and the like. The future is exciting…