The Future of Robots Doesn’t Look Like What You Think It Looks Like

Our machines are very smart. They’re also very incompetent.
Author:
Publish date:
Social count:
0
Our machines are very smart. They’re also very incompetent.
Glenngarry Glen Bot. (Illustration: Susie Cagle)

Glenngarry Glen Bot. (Illustration: Susie Cagle)

We cannot wait for the future. Our impatience is often so acute that we convince ourselves that the future is much closer than it truly is. We excel at hope, but we are, in general, quite awful at predictions.

We have been predicting that machines would take over all our jobs since the Industrial Revolution, when machines did, in fact, take over many people’s jobs. In early 2014, researchers Carl Benedikt Frey and Michael A. Osborne wrote a widely cited paper predicting that 47 percent of all jobs were likely to be automated within the next “decade or two.” The number is sold as both worker empowerment and company cost-cutting, but, in many ways, it’s neither.

It’s far easier to build a machine that can design a building or calculate the risk and reward of a financial investment than it is to build one that can fold laundry.

Manual, physical labor has traditionally been the most susceptible to innovation, and Frey and Osborne posit that many of the remaining manufacturing jobs will soon disappear, along with many in office administration, telemarketing, and transportation. This may seem especially disturbing in light of an incident at a Volkswagen factory in Germany earlier this week, where a robot “grabbed and crushed [a worker] against a metal plate,” killing him.

Automation is best suited to the kinds of routine-heavy, middle-wage jobs that are already on the decline in the United States, as the economy diverges further and further into two opposing extremes.

We don’t want this. We are much more comfortable with the idea of robots cleaning our floors and fetching our coffee, the kind of work we value the least, the kind we already often treat as inhuman. Frey and Osborne say that service is one of the sectors most likely to soon be dominated by machines.

But robots are really bad at those things. Robots can’t walk up stairs. Robots can’t pick up cups. Robots don’t understand what we’re saying. Robots can’t even dump this bucket of Lego heads into a chute, and then scoop them into another chute.

And yes, they can kill you if something goes wrong.

It’s far easier to build a machine that can design a building or calculate the risk and reward of a financial investment than it is to build one that can fold laundry. Artificial intelligence algorithms have already taken over many of those high-level tasks. So why is our financial industry still so bloated and so well-paid? Operating at peak efficiency would look like way fewer bankers, but whose interests would that serve? When engineers create a robot that can engineer other robots, will they lay themselves off and declare victory?

But there is also a danger in obsessing too much over the future of work: It allows us to ignore the very present reality of work, and the economics that dictate it.

There’s a school of thought that believes automation will free society from the shackles of overwork and force government to institute a strong social safety net, including a universal basic income. This is my favorite vision of the future, but I fear it, too, is a fantasy. As the middle class shrinks, more are pushed into low-skill, low-wage jobs that may or may not be imminently roboticized. If and when they are, what will force the government to finally develop a compassion and a large budget for the unemployable?

We can decide to make our world more conducive to menial robot work. We can outfit our lived experience with affordances that make that work far easier—barcodes and zipper tabs, for example. But deciding to create a more equitable economic system will be far more difficult. Even though sustained assistance has often proven cheaper than addressing problems after they’ve become entrenched, humans programmed by capitalism seem unlikely to undermine capitalism.

We are currently a little obsessed with “the future of work.” Predicting it could help us prepare, learn new skills, and better ensure our survival. But there is also a danger in obsessing too much over the future of work: It allows us to ignore the very present reality of work, and the economics that dictate it. That Volkswagen factory robot seems terrifying, until you consider that about 21 workers are killed each year by being caught or crushed in machinery in the U.S. alone. This is not a problem posed by the future, but a reality of industrial labor that has existed for more than a century.

But Carl Benedikt Frey and Michael A. Osborne can rest assured. One field that is very unlikely to be automated any time soon is that of predicting the technical future.

The Crooked Valley is an illustrated series exploring the systems of privilege and inequality that perpetuate tech's culture of bad ideas.

Related