Book Review: Acemoglu, D., & Johnson, S., (2023), Power and Progress, Basic Books

In Power and Progress, MIT economists and Nobel Prize winners Acemoglu and Johnson argue that, when analyzed through economic history, technological advances will not automatically lead to broad-based prosperity, but instead often end up benefiting only the wealthy elite. Will AI, for instance, increase the average worker’s productivity, or will it simply create a more exploitative workplace run by very few at the top? The two authors demonstrate that although technological advances most often tend to benefit only those at the very top, we can steer technology to promote the public good. We can indeed take control, but this will largely depend on the small changes that we must make. To focus on this is particularly urgent since digital technology and AI seem to increase inequality and undermine democracy.

The authors argue that we cannot stop technological change, but we should shape it!

When it comes to having control over technology, Steve Jobs famously said: “Let’s go and invent tomorrow, rather than worrying about yesterday”. But, who won, who lost, and why? It is a struggle to come up with the “answers”. An initial discussion of this issue is the successful development of the Suez Canal, led by Ferdinand de Lesseps, as well as the significantly more difficult project of constructing the Panama Canal. It was the investors who benefited, not the thousands of workers. Salaries did not rise, and adverse health conditions became rampant. Also, while de Lesseps had a clear vision, it became unrealistic and distorted.

Switching now to the more recent economic crisis, the ability to set agendas seems to be particularly critical. The more powerful people become, the more likely they are to act selfishly, and ignore the consequences of their actions on others. Their power to persuade becomes corruptive.

To find ways to curtail such selfishness, the authors suggest a strong, “live” democratic process. Here it seems to be particularly important that a broad strata of people are well informed and politically active.

The authors now refocus on long past economic history - the Middle Ages - where they examine on innovations in the agricultural sector. While innovation was significant, there were very few benefits to the farmers. Most of the proceeds went to the construction of chapels and cloisters. Persecution, and the idea that this was the will of God, was center stage. The elite chose the path for agricultural innovation based on their own vision. Ordinary people paid the price!

This pattern basically continued during the industrial revolution. The benefits of innovation were to a relatively small group of people at the very top. To solve practical problems was perhaps a key to the progress, i.e., a focus on implementation, rather than more basic science. Wealth was also important. With wealth in hand, there seemed to be few limits regarding how high one could rise socially!

But industrialization also brought a great deal of pollution, particularly CO₂ emissions. This was a consequence of a large increase in the use of coal. The UK, for instance, had become a deeply unhealthy nation. The biases due to an excessive focus on technology led to reactions such as emergence of trade unions.

While many similar problems from rapid industrialization also emerged in the US, there seemed to be two main factors that steered the innovative energy direction positively- mass production and a systems approach. The availability of electricity was central to this. The Ford Motor Company is a prominent example of how electricity, engineering, the systems approach and new tasks came about. Labor movements seemed to focus on reducing costs, with an effect being that the resulting economic benefits were somewhat shared. But nevertheless, economic inequality developed further with excessively wealthy owners at the top. Many significant worker groups were basically excluded: women, minorities and immigrants. The authors briefly discuss how some of these dysfunctional factors seemed to be handled better in various other countries, such as Scandinavia.

This brings us to what this reviewer sees as the central part of the book, namely discussions of various aspects of digital economy, starting with digital damage. While some sources, perhaps most notably Milton Friedman, claimed that the primary responsibility of business was to increase its profits (the market then self-regulates), there were others, such as Nader, who called for various forms of regulation in society. But big business grew! Sticking to what was actually working seemed to be critical. New inventions were actually held back, so as not to undermine what actually worked. Cost cutting, also spearheaded through the introduction of main-frame computers and new office software, became the norm. The US tax system forced capital, in contrast to working people. Robots were now also introduced on a large scale, while many jobs were “exported”. There was an emerging social bias of technology, with an antilabor bias of automation, in particular.

Norbert Wiener, the world-famous MIT-based mathematician who was considered a genius, said as early as 1949: “The world of the future will be an even more demanding struggle against limitations of our intelligence, not a comfortable hammock in which we can lie down and be waited upon by our robot slaves”. Fears of dysfunctional effects from AI seem to underscore what Wiener was thinking about, in particular. The authors point out, however, that human adaptability and ingenuity are less important today than in the past. Technology-minded executives may seem to ignore this, with AI seemingly having a lack of social intelligence!

We also see that digital technology can be used to come up with more effective ways of monitoring us humans, such as, for instance, the wide use of face recognition in China. Going one step further, the spyware package, Pegasus, (developed by the Israel-based firm NSO) seems to be in wide use— not only in totalitarian countries, but also in the US and Europe. We see politics-driven excesses, such as hate speech, particularly on Facebook. So-called filter bubbles underscore that people generally prefer to hear messages that are aligned with their political views. Much social media is driven by advertising, Google in particular. In fact, however, social media is becoming a cesspool! Is there an anti-democratic turn? The authors definitely seem to think that AI might be undermining our democracies when we need it the most.

The authors then discuss how technology might be redirected. They come up with a three-step approach:

-       First, attempts to change the norms. For instance, to curtail CO₂ emissions through various forms of filtering and storage used to be a central theme. But now, “new” ways of generating electricity (solar, wind, hydro) have become less expensive than conventional fossil-fuel based electric power generation, i.e., a redirection of technology.

-       Second, cultivating countervailing powers, so as to “remake” digital technologies. Such countervailing processes may be focused on automation, surveillance, data collection, and/or advertising.

-       Third, to come up with new policy solutions, for instance, for improving the productivity of workers in their current jobs, creating new tasks through AI to augment human capacities, providing better information for human decision-making, and/or developing more platforms that bring people together with different skills.

Countervailing powers may be built in different ways. The authors suggest three:

-       New methods for organizing workers.

-       Societal actions, such as systems for electric cars (battery technology, electric motors, charging stations, servicing, etc.).

-       Online meetings and communication, which might mean faster communication than before, “global” teams, less air travel, etc.

Closely linked with these factors are eight other policies that the authors propose for redirecting technology:

-       Develop more effective incentives for redirection, such as improved management techniques for assuring impact.

-       Breaking up “big” technology firms, such as Google, Microsoft and Tesla.

-       Tax reform: lessen taxes linked to salaries, and (modestly) increase taxes on capital invested.

-       Invest more in training programs.

-       More active government leadership to redirect technology away from today’s “laissez faire”. What the government did by proactively taking action to speed up the development of an effective vaccine for COVID might be an example.

-       Protect privacy when it comes to data collection.

-       Repealing the 1996 US Privacy Protection Act, which basically allows internet platforms to operate without much restriction.

-       Impose a (small) digital advertising tax, and also limit the dependence of corporate sponsors.

In summary, the authors have come up with a unique way to look at technological progress, built on the following four pillars:

-       How the benefits from improved technology should also be shared with those who are working within an organization, and not primarily the owners only.

-       How choices should be made, so as to direct technology to provide more societal benefits.

-       Better “mechanisms” for sharing the benefits of new technologies among all: salaries, safety, …

-       Bring non-economic factors more to the forefront when it comes to technological change – above all, societal and political power elements.

This reviewer finds it particularly compelling that lessons from history are drawn on. Automation, for instance, can, but does not necessarily reduce wages. Large income inequalities seem to make it relatively more difficult to create fairer ways for sharing the new benefits that are now arising. The politicians in our societies do seem to play important roles when steering the direction of future technologies. It might perhaps be a call for somewhat more activist government sectors in our societies, such as those we are seeing in Scandinavian countries, i.e., a rebalancing away from total “laissez faire”, but not totalitarianism.

This book is on the top of my reading list, and is even more important today than when it was initially published in 2023. While it represents a wake-up call when coping with today’s main technological challenges, such as AI, this book also provides a clear blueprint for how to handle these challenges!