This morning, Joe Paduda, at Managed Care Matters, did the workers’ compensation community a service with his blog post What Job?, in which he writes about the likely prospect of many jobs disappearing due to Artificial Intelligence (AI).
AI is permeating every industry. For example, watch Motoman’s nearly human-free plant in Sweden as robots assemble, package, shrink wrap and ship thousands of Ikea bookcases, or a BMW plant in Germany as robots build the i3 electric car from scratch and with amazing precision, or CNN as its reporters show how Google’s driverless cars are navigating California and three other states.
This is one of the defining issues of our time, and one in which I am profoundly interested. I suggest it is not hyperbole to predict that we are on the verge of an epochal change, something like a kind of mass extinction, and what’s going extinct is an enormous number of jobs. This change might be even more significant than humanity’s evolution from an agrarian to an industrial economy.
Mr. Paduda cites the the Frey and Osborne paper from Oxford University’s Oxford Martin School that suggests up to 47% of jobs run the risk of going the way of the Woolly Mammoth by 2025. The paper is highly controversial and its conclusions have been hotly debated. Even so, the most conservative naysayers agree that the figure is at least 16% (but that figure doesn’t take into account the jobs that the AI revolution will create, estimated at about 9%, for a net job loss of 7%). So, you can see there’s a huge difference of opinion. Regardless, millions of jobs will be lost. Add into that mix the following, and you have a wicked brew for the future:
- The rise of the Millennials, whose cultural and social thinking, as well as their total buy-in to AI technological development, is nearly diametrically opposed to today’s economic leaders, the Baby Boomers, who are now exiting the work force at the rate of one every seven seconds;
- Rapidly increasing economic inequality throughout the workforce. For example, consider that since 1973 inflation has grown 533%, while hourly wages for non-farm US Production and Non-Supervisory employees, measured in constant 1984 dollars, have risen a whopping 4%;
- Put that beside the Disappearing Act of US labor unions. Whatever you think of them, unions were good at negotiating hourly wages and other benefits like health care. But with only 6.7% of the private sector unionized today, that doesn’t happen anymore, so we’re left with a near total change in how we pay for health care today, as well as what we get for it, compared with how we paid and what we got for it in 1973. For example, we now have $2,000 (or more) deductibles, significant increases in employee premium contributions, the gold medal for the highest cost for health care in the developed world, way-over-the-top histrionics surrounding the Affordable Care Act, etc. You get the picture.
- The increasing need for both cognitive and social skills in entrance level jobs on track for middle class wages (presuming there is a middle class). See The Increasing Complementarity of Cognitive and Social Skills, by Catherine J. Weinberger, University of California, Santa Barbara.
- The continued accuracy of the 1965 technological development prediction, known as Moore’s Law. Simply stated, Gordon Moore, the founder of Intel, predicted that computing power would approximately double every two years, and it has – since 1965. Moore has recently revised his prediction to two and a half years through the next decade.
- Finally, the rise of what can be called the “intellectual” robots as exemplified by IBM’s Watson, which, in addition to handily beating all-time champion Ken Jennings at Jeopardy, is already doing a better job of reading XRays, CT Scans and MRIs than human radiologists. And here’s something to ponder along that line: Many agree that a workers’ compensation claims adjuster with a lost time caseload of 100 to 125 claims has the ideal caseload (a few nights ago at a Central New Jersey Claims Association meeting I was talking with a motivated and energetic workers’ comp claims adjuster for a company that shall go nameless – but it’s respected and well-known – who told me she’s carrying 180, and that’s the norm for her group). How would Watson do on that? Watson, a machine that sounds like the Hal 9000, never sleeps and could analyze thousands, maybe more, of lost time claims a second. IBM says that human involvement with Watson is critically important, but how many people does that mean? After the initial intake with the claimant, how much claim management could shift to Watson? You may not agree, but it seems to me that more than a few workers’ comp claims adjusters may soon begin to resemble that long-gone Woolly Mammoth.
I’ll leave you with a question: Do you think it would have been appropriate if this subject had made it onto the agenda for the upcoming WCRI Annual Conference in Boston, March 10 and 11? Regardless of your answer, I’ll be there and will be happy to dive deep into the AI crease with any interested attendees.
Hope to see you there.