Automation panic, then and now
Artificial intelligence is replaying an old scare. But the threat to jobs is real.
Pew Research Center has a new survey which shows that “while 73% of AI experts surveyed say AI will have a very or somewhat positive impact on how people do their jobs over the next 20 years, that share drops to 23% among U.S. adults.” Women, the survey showed, are way more pessimistic across the expert class and among regular folks. For example, 22% of men think AI will positively impact the U.S., compared with 12% of women.
If you expose yourself to enough AI panic and anxiety, from Hollywood to manufacturing, you might be justified in your pessimism. Here’s a too typical headline: Not yet panicking about AI? You should be – there’s little time left to rein it in. Another one: AI anxiety: The workers who fear losing their jobs to artificial intelligence.
Much of this is a replay of the panic and anxiety over factory automation that welled up in post-World War II America. Back then, automated processes of all kinds got the attention of not only of economic theorists, but Congress and the White House and the biggest labor unions in America. Industrial automation was viewed, correctly as it turned out, as revolutionary, just as AI is predicted to be today. In the following video, the UAW’s Walter Reuther took a realist approach, understanding that boosting productivity was good for the economy yet, he cautioned, living standards had to keep pace. Machines don’t buy automobiles, he reminded the Congressmen.
Mr. Reuther goes to Washington (mid 1950s):
Reuther:
We don't object to that expansion in our productive capacity because we know the workers that I represent they're grown people they know that there are no economic Santa Clauses that you can't get something for nothing that you can only have higher living standards and more of the good things of life only as you create the economic wealth by the application of human labor and human itelligence to the tools of economic abundance and the economic access that we have access to so they say fine we've got a bigger plant capacity we've got more modern and more productive tools.
But they say if we aren't using these tools and we've got idle capacity and unemployment then why and the answer is we aren't using our productive capacity because we haven't got the purchasing power in the hands of millions of families necessary to translate human needs into demand on the store counters of America and that's why we're in trouble.
Unions were under no illusion about job losses, principally the unskilled and semi-skilled. Reuther noted this in his remarks to Congress. In a research paper on unions and automation, scholar Adam Parker looked at how labor organizatins came to grips early on with the automation revolution.
… the AFL-CIO and its industrial unions were clear that they supported automation because of the benefits that it would have for its members and for society broadly. Reuther wrote in 1954 that, with union guidance, automation could “bring health and happiness, security and leisure, and peace and freedom to mankind everywhere” (UAW-CIO Ed. Dept. 1955). The UAW further felt that because of automation “many of the unpleasant jobs will be eliminated”.
Reflecting a common view, Abraham Weiss (1957) of the Brotherhood of Teamsters claimed that “automation removes routine and uninteresting work” and that it “removes human drudgery”. Automation was also thought to increase the skill levels required of workers. In testimony before Congress, Reuther noted the “general agreement that one of the results of automation will be a substantial raising of the level of skills required in automated factories and offices” (1955).
Parker’s paper is worth reading in its entirety for his analysis of the “divergent responses of different unions to automation.” A lot of that had to do with what type of occupation that was represented in a union’s rank and file.
Divergent would be a good way to describe the approach of the International Longshoremen’s Association (ILA) when it won record wage increases and safeguards against automation earlier this year. The union credited President Donald J. Trump for his support. Earlier President Joe Biden said he wouldn’t invoke federal law to suspend the strike if the longshoreman shut down ports on the East Coast and the Gulf. Maybe just as important, the Teamsters union promised not to cross the longshoremen’s picket lines.
ILA President Harold J. Daggett said his union “is steadfastly against any form of automation — full or semi- —that replaces jobs or historical work functions. We will not accept the loss of work and livelihood for our members due to automation. Our position is clear: the preservation of jobs and historical work functions is non-negotiable.”
For now. As I’ve said here before, if you belong to a union that doesn’t have the power to cause serious economic pain to an employer, it’s not much of a union. It’s just a dues collecting mechanism to support the lifestyles of whomever is running your union. The ILA has the power to shut down a good swath of the U.S. economy and it was prepared to use that power. Everyone knew that. The ILA won big.
Bu you have to wonder how long the ILA’s bare-knuckles approach can withstand the rapid advances in AI and robotics. Tell me where AI will be in six months or a year from now. And what about economic and political developments that no one can predict. Trade war anyone?
Reuther, some 70 years ago, observed that the “only way that we can have more is to make more and the only way to make more is to have more productive tools.” And that’s exactly what has happened in auto plants.
If you’d like a glimpse of how even as late as 1970 much of the work in an auto plant was “unpleasant” sample the “Golden Era” video below. A lot of manual processes. You’ll also see early digitization of design and toolmaking — and a lot of punch tape.
At about the 9:30 mark they show “tack welding” in a body shop where the process is already partly automated. I remember watching friends of mine tack weld bare metal auto bodies in the summer heat at a Fisher Body plant in Pontiac: hot, dirty and dangerous. Not much of a golden era, if you ask me. Robots do the welding now.
Yes, many of these factory tasks were eliminated or automated and no one should mourn that.
Which is how American manufacturing has become steadily more productive, although manufacturing employment has plummeted. Consider (yes, AI generated bullet points):
Manufacturing Output vs. Productivity in the US (1975-2025:
Manufacturing output has grown significantly despite common perceptions of industrial decline. The US manufactures about 2.5 times more goods today than in the 1970s in real terms.
Manufacturing productivity (output per hour worked) has increased dramatically, growing at a faster rate than in most other economic sectors.
Manufacturing employment has declined substantially, falling from about 19 million workers in the 1970s to approximately 12-13 million before the pandemic.
Automation drives productivity.
I know that abstract talking points about productivity are cold comfort to the cities and people left behind when old and obsolete plants are shuttered. Or when perfectly good plants pack up and leave for low-wage havens in China and Mexico and elsewhere. Michigan is littered with these abandoned and once highly productive factories.
Here’s an idea. Create a more congenial business environment for manufacturing by lowering taxes, paring back ridiculous and job killing regulations, and making sure plenty of affordable energy is on tap. Do all this without any taxpayer bribes to companies for doing what they should be doing anyway. Make it easy for companies bringing production back to the United States to locate in those places that have once been abandoned. Find ways to promote worker upskilling and trade education. But no worker training unless there’s a job to be had.
Reuther was right. Machines don’t buy automobiles. And, we can add, robots don’t pay union dues.