Visiting family recently in Silicon Valley, and driving up and down Route 101, I was struck by how many billboards one sees touting the benefits available to firms who use the latest internet technology. These are not specific ads for cars, or health products, or chain restaurants, but promotions for advice-giving firms – e.g. Oracle – which knows how to predict trends, improve sales techniques, identify what customers want and how to enlarge markets (called ‘customer relations management’ – CRM), and other (lawful) means for raising corporate incomes.
At another, even higher level of theorizing, firms like Brighterion offer products and services based on concepts of “Artificial Intelligence” (AI) and “Machine Learning”. Being of philosophical temperament, I find this trend in technology particularly intriguing. What does the term AI actually mean, and how will its use in the economy add to trends that are already showing the pervasive and overwhelming influence – indeed the addictive nature – of the internet in everyday lives?
I’ll leave the philosophical critique about whether, and to what degree, machines can properly be called ‘intelligent’ for another discussion. But in this article we’ll see some of the controversy over what the trends in IT technology may bring about – specifically in terms of the impact, for good or ill, of the increasing use of ‘robots’ in all sorts of industries. Whether or not the new tools constitute intelligence is a moot question. They do indeed have powerful effects on the way we live day to day, and the pace of change is increasing. I don’t think it should all be called progress, however, either in itself, or in its consequences. A major question is beginning to be talked about in this regard. Will the new technology in robotics leave more workers displaced, within a decade or two, than the jobs that result from it?
A quick Google search (yes, I know …) suggests that these developments are indeed going to be socially damaging, harming the lower paid workers especially, and increasing inequality. There wasn’t a single source I could find that expressed the view that robotics will, in the near future, lead to as many jobs as it eliminates, although that is the historical pattern for technological growth. Despite this agreement, however, I’m skeptical of the commonly accepted opinion that robotics is a major threat. What is the reasoning behind popular views? One sees that the major sources of opinions about the question – e.g. Washington Post, N.Y. Times, CNN, MSNBC, Slate, Wired, Huffington Post, Fast Company, and PwC (Pricewaterhouse Coopers, an internationally dominant, privately owned investment management enterprise based in London) – are all large corporations close to the center of the financial/ government/ media mix. They are prejudiced to be ‘believers’ in the IT juggernaut and its inevitable destructive power.
But why would they express these views openly? One might expect them to deny or disguise the socially damaging effects, or hide it from the increasing number of low paid ordinary workers today, and from older employees whose wages have been stagnant or declining for the past 40 years. But I think it’s actually to the advantage of the key industries and their media and political supporters to make this dire prediction (which they probably believe), because whatever effects it might have on the ordinary public, such technological advances promise to increase the productivity of their corporations, lower the costs of labor, and advance their investment potential.
It’s natural for a person to believe that what she would like to be true is true. So, because it would indicate a rosy future for the leaders in the affected industries, they are inclined to accept as fact the threat of the takeover of robotic technology. Some of them have even set about already to offer solutions to a problem that so far is only a supposition. The ‘solutions’ offered are ways of appeasing the displaced workers – e.g. by proposing a guaranteed minimum income, or by suggesting new kinds of job training, in skills that required sensitivity, sympathy and empathy (which robots can’t match, yet). Despite these suggestions, it’s likely that things will continue to change as they have been, but perhaps faster. Not only low-skilled workers, but most employees at various levels of education and income, will continue their downward glide in the ‘new economy’. This is what Michael Hudson calls the Slow Crash, which is in both text and audio. (I recommend the text transcription, which can be read more quickly, and was edited to be clearer.)
Even BloombergMarkets, whose analyses are more nuanced, and cite reputable academic sources, gives a pessimistic outlook. If you can, read the text part of the following page at BloombergMarkets, by Jianna Smialek, before viewing the accompanying, mind-blowing video with Maurice Conti, of Autodesk, showing how robots are teaching themselves to “see” their human co-workers, and be careful not to crush them when they do dumb stuff : “Professional Development for Your Robot Replacements“. My granddaughter was recently hired by Autodesk, to work in materials science in the 3D printing research project, located on the San Francisco Embarcadero. Here is a new prototype printer being used in tests at their “Studio”.
If all goes as expected, this device will soon be able to ‘print’ solid metal objects (like my newly replaced hip). There’s an atmosphere of creativity and enthusiasm among the employees in the multi-departmental ‘Studio’ here that is palpable. It’s easy to see how bright the future appears in this part of the economy.
Unorthodox views of the robot question are given in recent articles posted at my favorite website for information about what’s happening in political economy, why it’s happening, and what can be known with reasonable confidence about trends and predictions (of which there is never any shortage): NakedCapitalism. In research, I try to keep in mind a principle of modern popular media that influences the quality of information they provide: Bad news is more entertaining than good news. Even ‘public media’ follow this money-making dictum. The posts I found especially helpful for understanding the robotics issue are Georgios Petropolis’ essay, “Do We Understand the Impact of Artificial Intelligence on Employment?” (April 29, 2017) and Kevin Cashman’s “The Automation Grift: From Flying Cars to Ordering Catfood on the Internet – Part 2” (April 22, 2017).
Petropoulous’ answer to the question whether robots will displace more jobs than they ultimately generate is, We can’t make a good prediction. That kind of response is not as popular as ‘Yes’ or ‘No’. He shows that historically, the use of ‘machines’ (including robots) has generally increased employment over time, though it has typically displaced individuals in the short run in particular affected industries. (That has certainly happened in ‘rust belt’ areas, and elsewhere, as the 2016 election cycle has shown.) For example, 19th century weaving factories in England made raw cloth products cheaper than by hand, which increased their sales at home and abroad, and more people could be well clothed and work in clothing-related markets. So the overall effect was economically positive. It was similar with the mass-production of cars, which developed whole new marketing niches and job opportunities (including travel agencies, motel businesses and fast-food chains), which no one would have imagined in advance.
But re-programmable robots (called “industrial robots” – i.e. the kind that can ‘learn’ multiple jobs) – may cause long term negative employment results, unlike in previous stages of technological advancement. They may, but we don’t know now. Most predictors are saying “This time it’s different”, but a more careful look suggests it’s a question that can’t easily be answered, despite the very many mainstream projections. The question is too dependent on variables that have not been researched, and historical patterns may not apply to today’s ‘revolution’ in ways that are relatable to past technological revolutions.
Petropoulos admits, of course, that major changes are occurring. “Thanks to complex virtual learning techniques, machines are now able to perform a wide range of physical and cognitive tasks. And the efficiency and accuracy of their work is expected to increase as AI systems advance through machine learning, big data and increased computational power.” The term “machine learning” doesn’t mean that machines can help to educate others (as in so-called “smart classrooms” and on-line teaching). It means that the machines themselves (i.e. – computers) can learn from their experience (the ‘feedback’ they get). This is aided by ‘big data’ together with more powerful computers.
The term AI (“Artificial Intelligence”) divides people into two classes, it seems to me. One group believes that all aspects of human thinking can, in theory, eventually be done by machines (generally called ‘computers’, although they do much more than mathematical computation). Moreover, they will eventually do it better than humans, since their memories and approaches can be linked and are cumulative in ways that surpass humans’ capacity.
The other group (including myself) believes that some kinds of human thought cannot be duplicated by machines, and never will be. Among these I would include imagining, expressing or interpreting tone (e.g. sarcasm, ridicule, condescension), feeling emotions (happiness, anxiety), being creative (art, poetry, music), and making moral judgments. I say these are beliefs, for both groups. The believers in technology (or science generally) will say, as they always have said about advances in scientific knowledge – ‘It hasn’t happened yet, but it can and will’. The disbelievers will say ‘It can never happen’. I think the believers are continuing what amounts to the much older modernist critique of religious thought: Religion is only an effort to answer questions that have yet to be answered by science.
Robot technology also relates to the larger movement in which religion has been battered by naturalistic thinking since the Enlightenment. (La Mettrie wrote his The Man Machine in 1747.) There is unquestionably a steady decline in affiliation with traditional religious denominations, which is greater in Europe than in America. That’s partly because, unlike in America, European history associates religion with royalty, class distinctions and state sponsorship. This religious decline is not, contrary to what might be expected, simply a result of greater knowledge about nature and its its laws. Lower levels of belief do not correlate with higher levels of education (at least among Christians, who are still the dominant religious group in this country). It’s safer to say first that ‘religiousness’ is becoming more oriented to ‘spirituality’, and that the role of religion in peoples’ lives is now more emotional than explanatory – especially in the sense of community, mutual support, and giving purpose to lives. Perhaps religion is “more durable than we thought”, as a recent NPR report suggests.
Returning to the question of AI, and Machine Learning, some view it as a major threat. Not only can it make people feel demeaned, or even superfluous. Some high-profile scientists have worried that these developments present an existential threat to humanity – an idea expressed by Stephen Hawking in 2014. Humans may be replaced, superceded, or even destroyed by self-guiding AI systems which find them to be in the way of progress.
Recently, a group of major players in ‘Machine Learning and AI’ have come together, sponsoring collaborative efforts to understand and address the development and social impact of this new aspect of technological change. It remains to be seen whether this will be self-serving PR, or work some societal benefit, e.g. by helping ordinary citizens, schools and policy-makers to understand what’s going on, and how to act appropriately. To see see who the big player sponsors are (the “partners”), and find what its goals and ‘organizational philosophy’ are, look around the site here. It’s worth noting that the 10 partners shown are not independent collaborators. Google has taken over DeepMind, while the non-profit OpenAI was founded, and promised $1B by Elon Musk (of Tesla fame), who has also invested in DeepMind. [See Wikipedia, “Deep Mind”.]
In discussions about AI and robots replacing workers, data seem to suggest that ‘ordinary’ jobs , i.e. low-level, low pay, service work, like farm laborer, landscaper, house cleaner, parking attendant, receptionist, checkout clerk, fast-food worker, table waiter, laundromat attendant, etc. are not being replaced. Here is a summary of that idea, from Kevin Cashman’s article, “The Automation Grift”:
This doesn’t mean that more original uses for technology couldn’t significant impact specific sectors. However, it’s likely that, in general, technology that does affect jobs will complement those positions, replacing or changing the specific tasks that workers do, but not going as far as replacing them in all cases. For jobs that are replaced wholesale, it shouldn’t be assumed that they will disappear overnight. There still need to be decisions, investment, and planning involved in replacing workers with (usually expensive) alternatives, which are all things that take time. This has certainly been the case in manufacturing. One interesting table from the Bureau of Labor Statistics that supports this point details the fastest declining occupations. Even extrapolating out ten years, the BLS assumes that there will be significant employment in these occupations. And any changes will vary by specific industry and occupation. Even then, many “low-skill” or low-paying jobs, especially in the service sector, are not conducive to automation very much at all. (And the robots must have forgotten that those were their targets, since many of the fastest growing jobs require no formal education or only a high school degree.)
In passing, Petropoulos mentions that roughly half of experts of one survey believe AI will be socially disruptive and displace more workers than it replaces, but half believe it won’t. I went to the 2014 survey he referred to, that was done by Pew researchers, of which the next paragraph is a summary. Note that they are careful to make clear that this was not a random sampling, so they have added a disclaimer at the end of what they labeled a “canvassing“. This is the Pew study summary:
Key Findings: The vast majority of respondents to the 2014 Future of the Internet canvassing anticipate that robotics and artificial intelligence will permeate wide segments of daily life by 2025, with huge implications for a range of industries such as health care, transport and logistics, customer service, and home maintenance. But even as they are largely consistent in their predictions for the evolution of technology itself, they are deeply divided on how advances in AI and robotics will impact the economic and employment picture over the next decade. We call this a canvassing because it is not a representative, randomized survey. Its findings emerge from an “opt in” invitation to experts who have been identified by researching those who are widely quoted as technology builders and analysts and those who have made insightful predictions to our previous queries about the future of the Internet. (For more details, please see the section “About this Canvassing of Experts.”)
To summarize the ‘thesis’ I am proposing: against the common fearful predictions about the dire effects of industrial robotics, it may be that outcomes to the economy will generally be more positive than negative, and that low wage service jobs in particular may actually be spared in the technological revolution involving AI and Machine Learning. This is because these outcomes are beneficial to financial interests. Financiers will continue to gain control of more assets and products, and inflate their prices – e.g. desirable land, scarce natural resources (like water), computer services and programs, patents on equipment, healthcare, pharmaceuticals and education, and to own the debt that burdens people at lower levels who must borrow more and more, to use those assets and products.
If lower-cost labor is ‘left alone’ – i.e. not driven to abject poverty – wage earners (and even salaried workers) will probably not rebel, especially as their political influence is minimal, although their skill levels, education, training and benefits will not be able to grow enough to improve their situation greatly.