“What technological progress has done to horses, it would eventually do to human beings as well: drive us out of work”
Wassily Leontief (Nobel Prize in Economics 1973)
“A World Without Work” is a book by Daniel Susskind, an economist (professor and researcher at Oxford University) who was an advisor to the British Government. He made himself known to a wider audience in 2018 when he co-authored, with his father Richard, the best-selling book “The Future of the Professions“.
Both books deal with a topic that is increasingly present in economic and political debates: the progressive replacement of the human being at work by a combination of robots and artificial intelligence systems.
Although there is no general consensus among academics or economic actors on this fact, Susskind fully subscribes to the idea.
He does so in an argumentative style that doesn’t fall into excessive alarmism or unavoidable destinies. It’s supported by facts and objective data that he presents to make us aware that this is the path taken. Perhaps he doesn’t prove it, but at least he tries to convince us with solid evidence.
Note: this is a ‘pre-pandemic‘ book published in January 2020, but I believe that this serious human and economic situation will only reinforce its message.
It wouldn’t be strange to see a greater commitment by companies to the process of automation/robotization in combination with teleworking (where feasible) to face this and future crises of this kind and ensure continuity of production and services in times of reduced mobility and social distancing.
I had already had the opportunity to read some books on this subject, such as “Only Human need apply” or “Rise of Robots” (see note at the end) and I decided to comment on this Susskind book as it’s the most recent one.
The book starts with a very illustrative story about horses and the impact that the invention of the combustion engine had on them.
It quotes the economist Wassily Leontief, who reflected this example in several of his essays in the 1980s: how an animal that played a central role in our economy (transport, loading, farm work…), was almost totally displaced from it in a few decades by the adoption of a new technology (which would also lead to a large reduction in its world population).
After this introduction he presents the main theme of the book: “Technological Unemployment“, a term that Keynes already used in the 1930s, although it wasn’t the first.
From that concept he addresses several aspects related to it that I will try to condense: it’s not an exhaustive list but I think it covers a substantial part of what is developed in the book:
- It explains what the introduction of technology into the economy and the world of work has meant, especially since the 19th century, in terms of:
- An explosion of productivity
- A huge growth of the economy
- A economic transformation due to new products, services and work profiles.
He calls them: productivity, “big-pie” and “changing-pie” effects.
Note: I have included a graph, (the book presents a similar one), showing the enormous increase in world GDP since the end of the 19th century and, specially, the second half of the 20th.
- It shows how technology plays two roles in terms of jobs: substitution and complementarity.
Technology can reduce human labour but also increase their ability to generate new products and services with greater added value, creating new employment profiles.
In the past, this creation of new jobs has outweighed the destruction of “old” ones, but the achievements of Artificial Intelligence, and its application in an increasingly wide range of tasks, have reversed this trend.
It presents us with the trends that can be observed in the 21st century: a reduction of jobs in the middle range of profiles and not so much in the high or low profiles, despite the fact that a greater impact was expected in the latter.
Discusses what we consider to be “Artificial Intelligence” (AI)
Personal note: this is not a book to learn about IA and its different technologies or categories (machine learning, deep learning…). Just an example of a comprehensive intro to AI at this link.
There’s a previous question: what is intelligence?
It’s a difficult concept to define. Some people end up identifying intelligence with those mental capacities of human beings that don’t yet have AI, that is, a shrinking area, perhaps confusing it with the elusive term “consciousness”.
It presents us with some interesting reflections such as the fact that AI does not necessarily have to carry out tasks or solve problems according to our methods, but can apply its own strategies that can be as effective -or even more- than ours. The key is what they can do and not if they do it following human algorithms.
- It describes which tasks are most suitable for IA to perform.
It analyzes a paradox: sometimes the tasks that require more intellectual capacity can be more easily automated and, instead, those that depend more on our motor-sensory capacity – for example, ‘manual’ jobs that are developed in changing conditions or those that especially require a human touch – are more difficult to carry out by a machine.
Therefore, it is necessary to reconsider which qualities of the tasks will make them more suitable to be carried out by AI.
(Personal note: until now, repetitive tasks with fixed rules that could be described in detail were very likely to be automated, but AI introduces flexibility thanks to its self-learning capability and this delimitation is no longer valid.
Factors such as the dependence of the physical environment or the context and if these can be changing or unpredictable, what data/information is handled (quality, quantity, standardization) or if there are defined objectives to be achieved or are somewhat diffuse, will be important
A couple of examples:
– Applying AI in a controlled laboratory environment is not the same as doing it in a real one: uneven results of medical diagnostic AI
– You will be surprised that nowadays it’s easier for AI to generate Bach style music than Pop style music, (the latter being more “simple” musically speaking)
- It raises the frictions that make it difficult to replace human work with “machine + AI”, such as low-wage jobs, legal restrictions or cultural factors. Therefore, Susskind doesn’t argue that this substitution is immediate or that it should be extended to the entire workforce, but he does expect it to be progressive and increasing.
- On the other hand, it raises something that not everyone agrees with: the limitations of education.
He sees an increasing difficulty in providing people with an education that enables them to compete for the jobs that the market demands, at least in the better paid jobs that require a high degree of training and specialization with constant updating of knowledge and skills.
Not all of us have the capacity (limitation in terms of vocations, time, skills, economic resources, effort…) to do so.
- Also warns us about the poor performance of training programs for senior workers that try to help them reorient their careers. At least that’s what some studies that have measured this impact show
- It warns us about the low unemployment rate being achieved in certain economies as the US: this doesn’t mean more jobs.
There is a fact that distorts these figures: there is a lower percentage of people looking for work, as happened in some periods of crisis, several decades ago. In addition, there is an increase in jobs with contracts for a very small number of hours.
Note: this is a pre-pandemic book, which doesn’t take into account the high job destruction that has taken place and that we will still have to check if it is temporary or more far-reaching.
Susskind also discusses the role he believes the State should play in the situation we are heading for, and he offers us his prescription, which logically will not satisfy all his readers depending on his political position.
In short, he expresses it very clearly: “State role is not production, but distribution“.
He proposes a “Big Welfare State” model as far as providing ample covers and guaranteeing vital minimums to all the members of the society, previous tax collection from different areas according to that budget.
(Personal note: since the book deals with applications of Artificial Intelligence, there are already projects investigating the use of AI to create more effective tax systems).
He rejects that State is the one “increasing the pie” because it’s less efficient than the private and personal initiative. In other words, the preference of the free market over directed, state-managed economy, but with a major redistributive role for the latter
- Another important point of the book is the need that the author sees to study and apply progressively a minimum income, analyzing its possible problems. The author lists a number of issues depending on whether it is a Universal Basic Income (UBI) or a Conditional Basic Income (CBI), only available to those who have no other income or are below a threshold. It prefers the latter and advocates linking it to the performance of beneficial community tasks that are currently unpaid.
- It points to the future and the role of “Big Tech” (the “Big Five”: Google, Facebook, Amazon, Apple and Microsoft).
He believes that this market concentration is more a political than an economic problem.
He finds it positive that their size allows them to bet and risk heavily on R&D and that they are not exercising an oligopoly, with illegally agreed-upon price controls since many of their services are free.
It is more concerned about the power they are acquiring through our data and the degree of control and influence in our lives (state of opinion, purchase decisions, knowledge of our preferences, movements, profiles…). Applying AI to this massive data management multiplies its information-based power.
The final part of the book is a reflection on the search for a purpose in life in a post-work society.
It reminds us that there were times when ‘work’ was considered as something that degraded the human being, more a question of slaves and servants. Not to mention jobs with almost subhuman conditions, from the factories of the early Industrial Revolution to certain current models of exploitation especially in developing countries.
This perception changed with the progressive stigmatization of the person without a job, considering him/her as a ‘scrounger’ of society, until arriving at modern and developed societies where the (job) work is considered a fundamental element in giving us a purpose to our lives and a status in that society.
In spite of this appreciation, we know that not few people see their jobs as a means of subsistence (or of income that allows them to do other things that are more interesting for them), not as something that gives meaning to their existence.
Susskind thinks that it is time to rethink this paradigm because there will not be jobs for everyone, but there can be tasks to be done (‘work’) for the benefit of the community even if they are not now paid.
He thinks that people should be able to choose in the future to have a ‘non-economic identity‘, equally appreciated by society and which gives value to their lives.
Well, after this tour through the content of the book, I would distinguish two issues in it: the presentation of the facts and the author’s reflections and analysis.
Just for its presentation of the current panorama, how it has evolved up to the present day and the possible scenarios that we are faced with, I think it’s already worth reading the book
As for the author’s opinions and proposed solutions, although I can agree with them to a large extent, I understand that there will be no consensus. However, the lack of sometimes fundamental details – it is not possible to develop the full extent of the issues presented in a book like this – prevents them from being properly evaluated.
I don’t think that’s a bad thing. The fact that the author sets out some prescriptions invites us to analyse his proposals and look for more answers among other different approaches in order to be able to contrast.
I also liked its realistic tone, moving away from the science-fiction scenarios, full of catastrophes that some books show. He doesn’t preach an overnight revolution: he knows that there is friction in the application of these technologies, but that the scope and impact they will have cannot be ignored.
In short, I consider it a highly recommendable book, written in a lively and enjoyable style, well documented, that informs and teaches almost without your noticing and that makes you think.
This is what I expect from an essay of this type and for me it has succeeded.
Note on the book “Rise of the robots“by Martin Ford.
Published in 2015, was a best-seller in its category (essays on economy-technology) and won several awards as book of the year by various prestigious publications.
I also liked it very much and I recommend it. It also deals with a similar range of topics as Susskind’s, and I think that with these two books you can get a good overview of this topic.