One sure sign that the age of AI is now upon us is the proliferation of articles purporting to know definitively which jobs are immune from automation and which are most susceptible. These articles are enjoyable, most likely due to the small dose of schadenfreude that comes from seeing that a profession that you felt had overcharged you in the past has its days numbered. Some even go so far as to assign a percentage of probability to the chances of automation down to the decimal point, giving the whole activity a sense of gravitas. Of course, if you read enough of these articles you quickly notice the wildly inconsistent predictions. One will have relegated financial advisors to the dustbin of history, while another declares that the only job left in 100 years will be advising robots on mutual fund decisions.
Dig a little deeper into the “methods” of these studies, and the moniker of the researcher starts to seem pretty generous. One NPR article used the question “do you need to come up with clever solutions?” as one of the eight criteria for establishing how automatable a profession was. By way of intellectual rigor, this question has more in common with resume fluff than the scientific method. The problem is the “researchers” can’t help but let their own bias get in the way. When it comes to a competition between robots and humans, it is hard for people to not implicitly root for the home team.
Ready or not, the revolution is upon us | Data Driven Investor
"We must develop a comprehensive global shared view of how technology is affecting our lives and reshaping the economic…
So much so that often the burden of proof is on the automated system to prove it is better than the human. The better question is why couldn’t a job be automated, not if it can be automated.
The uncomfortable reality that all these studies choose to dance around is that there is no such thing as a profession that is immune to the effects of the automation.
The truth is if you can train a person to do a job, you can train an AI to do the same job- cheaper, better, and faster.
Furthermore, even if the whole job can’t be automated, most likely at least part of the job can be. After all, what is a job but a series of small tasks? A medical assistant might perform dozens of different tasks that each represents varying degrees of difficulty in automating, but over half their time is generally spent gathering information from patients, which is a job much better suited for AI than a person. Not only is it more cost effective, but studies show people give more accurate answers to a machine than a person, as they innately fear the judgment of others. By comparison, we don’t think twice about Googling the most embarrassing questions, provided we are in incognito mode on our browser.
Anyone who might be feeling smug about the difficulty involved in automating certain aspects of their job would do well to consider how a housing market crash occurs. In 2008, even in the worst hit states like Nevada, only about 7% of total housing was in foreclosure. However, that small minority of total homes still managed to drag prices of all homes down by over 40%. The point is, even if only 7% of a profession is automated it has the potential to send the job market in that industry into freefall.
Still unconvinced that your job is at any risk from automation? You feel it is so abstract and artistic that your spreadsheets are more at home in a museum than an inbox. Well unfortunately for you, your profession doesn’t exist in a bubble, so even if your hubris proves to be true and not a single percent of efficiency can be gleaned from adding AI to your job, it will still impact jobs that are lateral to your skillset.
The bad news is that workers that are automated out of other jobs won’t simply have the decency to take a hint and fall off the face of the Earth. Rather, they will start looking for new opportunities in less automatable professions, not unlike your job.
So much for best case scenarios.
Ultimately, we need to stop deluding ourselves that the proliferation of AI will create as many jobs as it automates. We need more people like Andrew Yang, ringing the alarm about the impact automation will have on eliminating jobs and start making plans for what a future with a lot less full-time work looks like. Finally, we need to realize that widescale automation is not a point for despair, but celebration. After all, most jobs suck, that’s why people are paid to do them. If long-haul truck driving and order fulfillment in an Amazon warehouse were such a hoot, they would be called hobbies and not jobs. We need to focus on saving people and not jobs.