In the wake of the 2016 Presidential election, we saw a battle among figures in what the late Molly Ivins called “the chattering classes” over the nature of Trump’s support. Nice people who want to think well of their fellow Americans identified economic insecurity, while not-so-nice others (including me) attributed the bulk of Trump votes to racism.
The ensuing research validated the racism connection, but of course, neither interpretation explained all votes or described all motives. It turned out that most Trump voters were not economically insecure, and researchers confirmed that “racial resentment” was the most robust predictor of Trump support, but there was one group for which economic insecurity was a motivating factor–prior Obama voters who switched to Trump. And the source of that insecurity was the steady increase in automation and AI–artificial intelligence.
Thomas Edsall reports on a recent study of –as he puts it–an “era in which vast swaths of the population are potentially vulnerable to the threat — or promise — of a Fourth Industrial Revolution.”
This revolution is driven by unprecedented levels of technological innovation as artificial intelligence joins forces with automation and takes aim not only at employment in what remains of the nation’s manufacturing heartland, but increasingly at the white collar, managerial and professional occupational structure.
The technological innovations we’ve experienced have ushered in an economy that rewards college-educated workers and disadvantages others, contributing to economic inequality. The scholars Edsall quotes predict that these advances in technology are likely to create additional social upheaval as they steadily affect the future of jobs.
Researchers find that exposure to automation correlates with support for Trump.
The strong association of 2016 Electoral College outcomes and state automation exposure very much suggests that the spread of workplace automation and associated worker anxiety about the future may have played some role in the Trump backlash and Republican appeals.
The study Edsall cites found that so-called “heartland states” like Indiana and Kentucky, both of which have heavy manufacturing histories and low educational attainment,
contain not only the nation’s highest employment-weighted automation risks, but also registered some of the widest Trump victory margins. By contrast, all but one of the states with the least exposure to automation, and possessing the highest levels of educational attainment, voted for Hillary Clinton.
That gets us back to the relationship between populism and automation. Edsall quotes an economist at Harvard’s Kennedy School, who studied those Obama-to-Trump voters.
Switchers to Trump are different both from Trump voters and from other Obama voters in identifiable respects related to social identity and views on the economy in particular. They differ from regular Trump voters in that they exhibit greater economic insecurity, do not associate themselves with an upper social class and they look favorably on financial regulation. They differ from others who voted for Obama in 2012 in that they exhibit greater racial hostility, more economic insecurity and more negative attitudes toward trade agreements and immigration.
In my last book, I addressed the threat automation poses to millions of jobs, and cautioned that humans tend to get meaning and purpose from employment. Edsall quotes from a 2017 paper in which economists Anton Korinek and Joseph E. Stiglitz went further, warning that artificial intelligence has the potential to create a high-tech dystopian future.
Without extraordinary interventions, Korinek and Stiglitz foresee two scenarios: both of which could have disastrous consequences:
In the first, “man and machine will merge, i.e., that humans will ‘enhance’ themselves with ever more advanced technology so that their physical and mental capabilities are increasingly determined by the state of the art in technology and A.I. rather than by traditional human biology.”
Unchecked, this “will lead to massive increases in human inequality,” they write, because intelligence is not distributed equally among humans and “if intelligence becomes a matter of ability‐to‐pay, it is conceivable that the wealthiest (enhanced) humans will become orders of magnitude more productive — ’more intelligent’ — than the unenhanced, leaving the majority of the population further and further behind.”
In the second scenario, “artificially intelligent entities will develop separately from humans, with their own objectives and behavior, aided by the intelligent machines.”
Unlike the Borg, Korinek and Stiglitz do not conclude that resistance to these possible consequences is futile. Instead, they advocate for government intervention and redistribution to counter the threats, leading Edsall to conclude with “the” question:
If fully enacted, could Biden’s $6 trillion-plus package of stimulus, infrastructure and social expenditure represent a preliminary step toward providing the social insurance and redistribution necessary to protect American workers from the threat of technological innovation? Can spending on this scale curb the resentment or heal the anguish over wrenching dislocations of race, culture and class?
I guess we’ll see.