Quincy Larson quotes Bill Gates, Elon Musk and Stephen Hawking as he warns
There’s a rising chorus of concern about how quickly robots are taking away human jobs.
He also cites an Oxford U. study
In 2013, policy makers largely ignored two Oxford economists who suggested that 45% of all US jobs could be automated away within the next 20 years. But today that sounds all but inevitable.
Gates, Musk and Hawking are the smartest humans in the world and I have cited them early and often in my innovation blogs and books. But, no disrespect, but none of them is a full time market watcher. We should take much more detailed studies from Oxford, Gartner (my former employer), McKinsey and other market watchers a bit more seriously.
Unless their studies are flawed.
In researching my recent book, Silicon Collar I found many flaws. Here are a couple in the Oxford study from my book
Nobody appears to have mapped the Oxford report to actual employment trends in the job categories they analyzed. The professors had calculated a high 0.79 “susceptibility to computerisation factor” (with 1.0 being the highest) to heavy truck and tractor-trailer drivers. This, when the U.S. trucking industry says driver shortages could reach as high as 175,000 positions by 2024 (even if the industry adopts autonomous trucks, regulations will likely require a driver as a backup). The professors had assigned an even higher factor of 0.84 to cartographers and photogrammetrists (who deduce measurements from images), which the BLS projects as one of the fastest growing occupations over the next decade. They had assigned a yet higher 0.94 factor to accountants and auditors, whereas hiring at U.S. public accounting firms jumped to reach record levels in 2013–2014.
Similarly, few appear to have asked the Oxford professors whether it is all doom and gloom. What about new jobs from the automation and new digital businesses? J.P. Gownder, an analyst at the research firm Forrester, is one of the few to have analyzed the Oxford work, and he estimated that “new automation will cause a net loss of only 9.1 million U.S. jobs by 2025.” His numbers are well under the roughly 70 million jobs that Frey and Osbourne believe to be in danger of vaporization.
Many of the jobs the Oxford study analyzed require combinations of finger dexterity, crawling capability, visual acumen, social grace, or cognitive and many other human skills. Of course, you can point to individual technologies that can match humans on each skill but show me a systems integrator which has put such a “Frankensoft” machine together for each of the 800 occupations the Bureau of Labor Statistics tracks? And when they do, how long before it matures? (before you answer, check what version of OS you are on after decades of evolution of the PC) And what will it cost to be competitive with a human?
In 2014, Gartner predicted “one in three jobs will be converted to software, robots and smart machines by 2025…New digital businesses require less labor; machines will make sense of data faster than humans can. By 2018, digital business will require 50% fewer business process workers.”
The fact is that Gartner issues hundreds of similar predictions each year and rarely audits them for future accuracy. It usually assigns a probability to them, as an indicator of its confidence in such a prediction, and this statement did not indicate any such hedge.
Gates, Musk, Hawking, the profs and the analysts and the rest of us also fall into what I call “automation amnesia”. We panic every 3–4 decades — because our generation has not experienced something similar in our lifetimes. The panic is often accompanied by other traumatic changes in our societies.
The current panic, I believe, was triggered by the scary, deep global recession of 2007–2008. No job seemed secure with the financial freefall we saw.
But if some of us had been in the workforce in the 60s, we would likely have remembered that President Lyndon Johnson set up a blue ribbon commission to explore growing panic about automation. The trauma back then could be explained by the turmoil around civil rights and Vietnam that the country experienced during his term.
Go back another few decades. Palo Alto, with its VCs and startups, is today the capital of the technology world. But would you believe the mayor of that city sent President Herbert Hoover a letter warning that industrial technology was a “Frankenstein monster” that was “devouring our civilization.”? The trauma back then came from the Great Depression of the 1930s.
You can go back every few decades all the way back to the Luddites and you find similar panic attacks. The Luddites, of course, had the ultimate panic attack. They were bands of English workers in the 1810s who destroyed newly introduced machinery, especially in cotton and woolen mills, fearing their jobs would be lost.
Studying the history of automation allows you to calmly assess what I call in this article “circuit breakers” that slow down societal adoption of automation
Agriculture has gradually been automated for centuries and we are still not done as we enter a phase of “precision agriculture” where GPS, drones and other technology allow for highly customized farm care. Ditto with manufacturing where we are moving to a new stage of robotics, wearables, 3D printing and other technology which is allowing cities like Greenville, SC to be reborn from a textile mill town to one which makes BMW SUVs for the world.
UPC scanners which started showing up in grocery stores in the 1970s have decades later not killed checkout clerk jobs. ATM machines and mobile banking have still not killed teller jobs (you may be shocked to hear we still have 90,000 bank branches with human employees in the US, and many times that around the world) Email and e-commerce may have reduced the demand for the delivery of letters, but they have not killed off the U.S. Postal Service. In fact, e-commerce has created an entirely new category of postal jobs related to delivering items ordered online. The robots at the mail marketing company Valpak and those at the distribution centers of Amazon and other companies help keep more than 600,000 postal employees busy.
I could go on.
I am not saying we should ignore the challenges automation poses to contemporary work and social arrangements. Millions of people will, in fact, find their jobs transformed by machines in the coming years. Automation typically targets tasks, especially 3D tasks — dull, dirty and dangerous ones. In doing so, it transforms jobs, rather than completely destroy them.
But in the short term, we should be far more worried about the threats posed to our labor economy by millions of unfilled jobs, massive student debt haunting younger workers, and an inadequate education system. But these wounds to the job economy are manmade. Don’t blame machines for that!
Cross posted at Medium
Book Review: What to Do When Machines Do Everything
As they say up front about likely impact of automation
It took me several pages in my recent book Silicon Collar to dispel the fears studies from Oxford U, Gartner and others have created about catastrophic job losses from automation. Frank and Co crisply do so with a “you have to be frigging kidding me!” section
They also bring out two important truisms about automation
a) Automation targets tasks, not complete jobs so it transforms jobs not eliminates them. (in my book I identified several 3D tasks – dull, dirty, dangerous – that automation has targeted. I also showcase examples of many grocery, banking, postal and many other jobs which persist decades after UPC scanners, ATM machines, kiosks and other automation targeted them)
b) we always seem to underestimate the “job gains”, many of which come from humble technology innovations. They use the example of the lawn mower without which today’s massive sports industry would not exist
The book is about more than automation - covers ground around digital transformation opportunities. In that sense the authors continue their tone from Code Halos. They nicely bring out the 3Ms - materials, machines and models -and their changing role across human progress in the last few centuries.
If I have one nit with the book - it focuses mainly on artificial intelligence (in its many manifestations) and white collar jobs. Automation in agriculture, the battlefield, logistics, oil and gas, manufacturing and many other sectors involve robotics, autonomous vehicles, wearables, 3D printing and other technologies. That’s certainly understandable. Cognizant is heavily focused on white collar industries like financial services.
If anything, that only reinforces how much further we have to go before systems integrators emerge who can blend a wide range of automation technologies to target finger dexterity, climbing capability, visual acuity, social grace, cognitive skills and so many other things the human mind and body are capable of but machines still struggle with.
Overall, an easy and pragmatic read - an important antidote to all the alarmist studies and books in the last few years on the topic of automation and the future of work.
February 16, 2017 in Industry Commentary, Silicon Collar | Permalink | Comments (0)