While I believe the future for humans is generally sunny, some workers face rainy days ahead.
This is unfortunate, but it’s largely beyond the ability of any one business leader or even one company to tackle. Businesses playing within the rules of the current system have no choice but to automate as much as they can.
Consider these two potential solutions for addressing negative worker disruption due to automation. Both require government intervention that goes beyond the scope of specific firms.
Change the Rules
One solution is to change the rules of the system to reverse or slow the pace of automation. Lawmakers could attempt to create a less competitive economic system; monopolies would have less incentive to automate because they have no competition to keep up with.
A more direct approach could be to target automation in specific industries: imagine a law mandating human drivers in all cars just as self-driving taxis are about to become ubiquitous. In the short term, this would obviously benefit the drivers on the cusp of displacement. But such an approach would harm society as a whole, especially when, as promised by self-driving car enthusiasts, the machine will do the job so much better than people can.
Redistribute the Gains
The second solution is to allow or encourage automation to proceed ahead while redistributing a portion of the benefits to those left behind. If, as I believe, automation, machine learning, and AI will yield tremendous economic benefits, then this is the better approach. For instance, imagine a machine that could bring $100 million of value to a community, but which would replace the work for ten people making $100,000 per year ($1 million of total value). Here is the economic loss incurred by outlawing the machine:
$1M in value retained
-$100M in opportunity cost lost
— — —
-$99M total value lost
Meanwhile, here’s what it looks like to put the machine in place:
$100M in value gained
-$1M in jobs lost
— — —
$99M total value gained
A small portion of this $99 million gain can compensate and retrain the ten workers who lost their jobs due to automation, leaving everyone better off.
The US government’s response to automation is a centerpiece issue for Andrew Yang, who is trying to become the Democratic candidate for President in the 2020 election. He proposes a universal basic income, “a type of social security that guarantees a certain amount of money to every citizen within a given governed population, without having to pass a test or fulfill a work requirement.”
Yang offers this solution because, his website says, “unlike with previous waves of automation, this time new jobs will not appear quickly enough in large enough numbers to make up for it.” It remains to be seen what legacy Yang’s candidacy will leave during this cycle, but I expect further political discussion of automation as machine learning and other technologies move forward.
While I’m more focused on machine learning than UBI and government policy, I do think business leaders should be aware of the broader socio-political context in which their own machine learning efforts take place. Businesses operate within a set of rules and assumptions about the marketplace. But as all leaders know, those can change.
Machine learning and automation will continue to transform businesses and employment opportunities for years to come. Executives have no excuse for being caught off guard as technology moves forward. Proactive companies will get ahead of the curve by cultivating positive mindsets from the C-suite to the factory floor.
It’s important to realize that the move to machine learning doesn’t mean that robots are coming to take everybody’s job. First of all, we’re talking about software, an image that hopefully brings the fear factor down a few notches. Second, the US job market appears to be doing fine. While we’ll eventually have another recession, we won’t see automation suddenly laying off half the country’s workforce.
For many workers, machine learning means software is coming to make your job better by automating tasks you’d rather not do anyway. For others, it means the chance for new kinds of work applying human ability to make machines work better.
Businesses, customers, and society overall are better off when machines, not humans, perform certain kinds of jobs. This economic transition, like others in history, will have a human cost. Addressing that cost isn’t a challenge for individual business leaders. It’s a challenge for society.
Robbie Allen is a Senior Advisor to Infinia ML, a team of data scientists, engineers, and business experts putting machine learning to work.