Reviewed in the United States on May 1, 2017
Martin Ford is among the technologists who have warned that technology is and will increasingly displace human beings across all employment sectors. This phenomenon will threaten the structural foundation of our capitalist economies fueled by consumer spending (that accounts for close to 70% of GDP in the US).
His vision regarding the exponential growth of technology (Big Data, Artificial Intelligence, Cloud computing, Internet and mobile apps, robots) is not all that different than the vision of Erik Brynjolfsson and Andrew McAfee as they expressed within their own two books on the subject: “Race Against The Machine” and “The Second Machine.”
Martin Ford puts more weight into what he calls the seven deadly trends:
1) Stagnant wages;
2) Declining share of GDP going to labor and rising share going to corporate profits;
3) Declining labor force participation rate;
4) Diminishing job creation, lengthening jobless recoveries, and soaring long-term unemployment;
5) Soaring inequality;
6) Declining incomes and underemployment for recent college graduates; and
7) Polarization and part-time jobs.
Brynjolfsson and McAfee also identified the same trends. It just that Martin Ford sees those trends as potentially irreversible meanwhile the two other authors do not. Consequently, how the
two sets of authors greatly differ is in their hope for the future of human labor.
Brynjolfsson and McAfee envision that humans have a fighting chance of eventually racing with the machine instead of against it. This means they view that ultimately digital automation could complement human labor instead of substituting it. They support their argument with an iconic example “freestyle chess.” In this new discipline teams of humans and computers team up in different ways to compete against each other. They have observed that the best freestyle chess teams are not stand-alone supercomputers who could easily beat Gary Kasparov; instead they are teams of mathematicians and engineers who have developed the rare skill of analyzing freestyle chess games while guiding simultaneously two or three laptops to run their adaptative Machine Learning algorithms. Brynjolfsson and McAfee advances that this represents a model for the labor force of the future on how to race with the machine. They advance that the collaboration between human intelligence and Artificial Intelligence (AI) will always be better (or at least for some time) than AI alone. By the same token they think that IBM Watson’s medical diagnostic abilities will be much improved when teamed up with expert doctors than on a stand-alone basis.
Martin Ford views things very differently. He considers the whole example of freestyle chess less than convincing even over the short term. He thinks that it is just a matter of time before AI renders humans superfluous. He sees no reason why AI would not quickly become better than humans at analyzing freestyle chess games. Similarly, he thinks IBM Watson on a stand-alone basis may not need the expertise of doctors to render superior medical diagnoses absent of human bias, flaws, and errors. Remember the entire work of Phillip Tetlock on the poor performance of experts across all fields because of overconfidence. Regarding medicine, the superiority of AI on a stand-alone basis has already been demonstrated with radiologists. AI was a lot more accurate and a heck of a lot faster at interpreting X-rays than expert radiologists were.
Given their positive outlook, Brynjolfsson and McAfee recommend we invest more in education focused on what they call “ideation.” The latter embodies intellectual curiosity, creativity, ability to come up with out of the box questions (some call it lateral thinking). They view those cognitive features as uniquely human.
Martin Ford, on his part, thinks that such investment in education may be good to maintain the social fiber of our society. But, it is entirely futile in terms of rendering humans able to race with the machines. Contrary to what Brynjolfsson and McAfee states, machines are already very good at many elements of ideation, creativity, exploring different solutions, etc. Machines can also be genuinely artistically creative. He gives an example whereby complex algorithms can now create wonderful classic symphonies that are competitive with the best composers. They can make modern sculptures. They can also write a surprisingly wide variety of news articles on numerous subjects. Someone in the book indicated that in a near future nearly 90% of news articles could be written by robots.
Martin Ford indicates that his concerns are not his alone. For a long time, leading minds have expressed such concerns. John Maynard Keynes stated back in 1930, that technology was likely to move faster than society’s ability to create new jobs for the displaced workers within just a few generations. In 1949, Norbert Wiener, a renowned MIT mathematician expressed similar concerns in a New York Times article. He stated “if we can do anything in a clear and intelligible way, we can do it by machine… [this will lead to] an industrial revolution of unmitigated cruelty [displacing human labor].” In 1964, a team of scientists and economists wrote a report for Lyndon Johnson called “The Triple Revolution.” One of those revolutions was titled the Cybernation Revolution. Within this report they reiterated the exact same concerns expressed by Keynes and Wiener. But, no one listened. Post WW II, the US economy was experiencing a long boom where technology created a rise in productivity, living standards, and overall economic growth at all levels (technology complemented labor and did not substitute for it). However, less than a decade after the Triple Revolution report was written in 1973 rising labor productivity became completely decoupled from stagnant wages and median household income. Brynjolfsson and McAfee also wrote extensively about this “decoupling” within their books and blogs. It is just that Martin Ford does not believe those decoupling trends can be easily reversed if at all.
When it comes to overall recommendations, besides education, investing in infrastructure, increasing Government investment in science and basic research, Brynjolfsson and McAfee recommend a modest increase in the Earned Income Tax Credit to counter the impact of technology displacement on consumer spending. Meanwhile, they are both adamantly opposed to a Basic Guaranteed Income for all because of moral hazard (incentive not to work) and the deleterious effect of idleness on individuals and society at large. While Martin Ford does agree with some public investment in education, infrastructure, science and basic research, contrary to Brynjolfsson and McAfee, he strongly supports the implementation of a massive Basic Guaranteed Income for all (qualified individuals) of at least $10,000. He views this as inevitable since otherwise consumer spending will get wiped out due to technological unemployment. He is not worried about the moral hazard of lowering individuals incentive to work since there will be not enough work to be had anyway.
Ultimately, Martin Ford in his analysis of the situation is much more convincing than Erik Brynjolfsson and Andrew McAfee. Yet I thought both their books (“Race Against The Machine” and “The Second Machine”) were formidable. But, they both claimed to be technology-optimists. Their positive recommendations were the weakest part of their respective books. Martin Ford, on his part, comes across as a technology-objectivist. He calls a spade a spade. And, he makes no artificial bias effort to dilute his diagnostic by making hopeful recommendations that he does not consider realistic.
There is just a remaining flaw in all these authors recommendation. And, that is that they all don’t fully factor or express the fiscal constraints (the US and other countries are under). The US fiscal position is already unsustainable. With the existing structure of social entitlements (very modest in scale vs. proposed Basic Guaranteed Income) the US prospective Budget Deficits are unsustainable. And, they lead to the US Debt/GDP ratio to nearly double from just under 80% currently to 150% by 2047 (CBO projection). Adding on an additional huge social entitlement (Basic Income) that would increase Federal Government spending by over 50% is just not realistic.
The above does not detract much from the overall quality of these outstanding books, and especially of this one book reviewed.