AI is going to take our jobs. Just not the ones you thought.
How to prepare for a world where machines do the math.
Open a browser, check your social feed, listen to the news and it’s impossible to miss the pundits pounding their drum: AI is coming, and — depending on which media guru you listen to — robots will take all our jobs and we’re doomed. Or, robots will free us from drudgery and the future looks bright. The techno fetishism and the techno dystopianism can’t both be true. Can they?
That divergence is an indicator that we need to pay attention to these developments and take a more nuanced approach. Technology can be confounding to understand. We don’t always know when we are talking to a bot, when we are engaged with AI, or when an algorithm (and not a person) serves us choices. We’ve already seen indicators for several years: in 2002’s editorial “If Tivo Thinks You’re Gay” in the Wall Street Journal and the true story of Target’s data mining program inadvertently outing a teenage girl’s pregnancy in 2012 before she could tell her parents.
But machine learning is really good at math — and pretty much only good at math. This means that the jobs that are really in the crosshairs are math-based. It’s the accounting and bookkeeping jobs that will go first, not truck driving.
Artificial Intelligence — which is just another name for Machine Learning technologies — is changing everything: it’s how autonomous cars run, medical algorithms diagnose disease and read x-rays, and Alexa pretends to understand you. Jobs are going to be lost. In the popular press taxi drivers and truck drivers are said to be most at risk. But machine learning is really good at math — and pretty much only good at math. This means that the jobs that are really in the crosshairs are math-based. It’s the accounting and bookkeeping jobs that will go first, not truck driving. Most of these jobs require only simple math and being up-to-date on policy classifications — and that’s exactly what ML is good at. These jobs are already being replaced by a new generation of accounting software like Xero (and, this, after so many more were replaced long ago by Intuit and Great Plains). These are job-killing machines. With Xero, you don’t need a bookkeeper, except maybe for the 10% of the time you require some advice or strategic thinking.
Insurance adjusters and underwriters are next and, really, anyone who has tried to get a mortgage recently already knows that it’s the algorithm that’s in charge, not your lending agent. Extend that to the IRS. Once ML automates tax filing that’s a lot of jobs going away. The only reason why the IRS doesn’t already fill out your taxes for you and simply tell you where to send the check (as most people in Europe experience) is that the lobbyists for tax preparers (I’m looking at you HR Block) make sure the IRS can’t be authorized to do the work for you.
And before you think you’re safe with that STEM-based coding or engineering job, guess what? Engineering is next. Autodesk has been baking the engineering into its design software for years, requiring less and less engineers to do the work once needed. Design a bridge and much of the engineering is automated. Architects should be worried, too. Complete a degree program, go to work for a large, prestigious company like HOM or Gensler and, the next thing you know, you are working on reflected ceiling plans for two years. Now, that incredibly unfulfilling work is going away, too, but without an associated boom in new architecture jobs to replace it.
At the precise moment when STEM people are shouting, “We need more math, coding, engineering, and software education,” ML is preparing to scorch the Earth of math-based jobs. Not all of them, of course, just most of them. The STEM message is all about quantitative optimization. And guess what: that’s exactly what AI and ML do best.
At the precise moment when STEM people are shouting, “We need more math, coding, engineering, and software education,” ML is preparing to scorch the Earth of math-based jobs.
So what’s left after the white collar jobs go? Jobs that require judgment in ambiguous or dissimilar situations — like plumbers and electricians. These jobs are secure. Also, jobs that involve a human’s touch: masseurs, beauticians, chefs. As it turns-out, blue collar jobs are safer than we thought (they may just pay better, too).
Success in the new world will require less math and more creativity. Preparation for the future requires innovation and critical thinking skills. These aren’t high on the STEM agenda — in fact, they’re mostly scorned. But, once AI and ML restructure the world of work, we will have to restructure education around creativity, innovation, and critical thinking. Instead of valuing education as career prep, we need to think of education as citizen prep.
“…Once AI and ML restructure the world of work, we will have to restructure education around creativity, innovation, and critical thinking. Instead of valuing education as career prep, we need to think of education as citizen prep.”
What we need are solutions that give humans a way to focus on the things they do well. Most people hate rote jobs already and many jobs people do now are simply not safe. So, why not give these over to machines? Humans have better things to do than plow fields, mine coal, watch endless live security camera feeds, or tabulate long lists of numbers. These aren’t actually our strong suits. Instead, we should be unleashed to use our critical and creative skills to solve problems, invent relevant, needed new things, and create.
For a long time now we’ve been building automated systems to do boring or dangerous work. Today, AI/ML technologies are poised to make our dreams of truly capable and even autonomous servants a reality. How will we control them? Conversation, of course — just as we’ve imagined for thousands of years. We used to call it magic or folklore or divine intervention. Now, it’s just technology.
The right platform can make this a reality. The wrong one could potentially make it an economic, technological, social, political, and human disaster. We need to start having these conversations now so we can start building the right solutions. We’ve started doing just that at Seed Vault and we’re looking for others to join us as partners.
Nathan Shedroff is the CEO of Seed Vault LTD, which is building the Seed Token project. A pioneer in the fields of experience design, interaction design, and information design, he is also the chair of the Design MBA programs in design strategy at California College of the Arts in San Francisco, and author of many books.