Can A Computer Do Your Job – And Does It Want To?

Doug Bernard | Washington DC

In 1949, writer Kurt Vonnegut saw something that amazed him.  Working at a General Electric plant, he noticed a large machine that cut the rotor blades for jet engines.  “This was a very expensive thing for a machinist to do, to cut what is essentially one of those Brancusi forms,” Vonnegut recalled years later in an interview with Playboy Magazine.  “So they had a computer-operated milling machine built to cut the blades, and I was fascinated by that.  This was in 1949 and the guys who were working on it were foreseeing all sorts of machines being run by little boxes and punched cards.”

The experience inspired Vonnegut’s first published novel, “Player Piano” – a bleak story set in a future where machines have replaced the working class, leaving them adrift and lost.

63 years later, while punch cards are as common as buggy whips, machines run by little boxes are in factories around the world.  These days, industrial robots are increasingly more nimble, accurate and reliable than even the most highly trained workers.  And the computers that control them – those little boxes – are moving beyond simple rote tasks and starting to demonstrate an ability to solve problems, thinking almost as a human might.

Still, many people probably feel as I do:  ‘But a robot could never do my job.’  Factory work is one thing (and I’ve worked in factories, operating machines like injection molding presses), but I’m a journalist – a writer.  I tell other people’s stories, hopefully with a little flair and lots of compassion.  How could a machine possibly do that?’

News flash: they can…and increasingly, they are.

Ned Ludd, Meet Watson

Since the beginning of the Industrial Revolution, there have been people who feared machines; specifically, that they would take their jobs.  “Ned Ludd” probably wasn’t the first to take hammer to machine – but he inspired many since to do likewise, even giving rise to the now-pejorative descriptor, “Luddite.”  Whether it’s John Henry of American folklore, the “steel-drivin’ man” who dueled a steam-powered hammer, or an auto worker’s union, fighting to keep more humans than machines on the production line, the battles are still with us.   As songwriter Maurice John Vaughn laments in his bluesy “Computer Took My Job“:

“I was at work that morning, when the big trucks came, with the big machines inside.  Computers.
My boss had us all gather round, he said ‘the computers gonna make your work easier, don’t you worry ’bout a thing.'”

You can see where that’s headed.  Which may explain, in part, why it’s so difficult for many to think about where technology is taking us.  Journalist Frank Tobe documents just that in his brilliant “Robot Report” – a daily collection of stories that track what robots are doing, and where they’re doing it.  Last week, Tobe took a deep look at the announcement that Asian tech giant Foxconn plans to deploy one million robots in its plants in just the next three years.

Foxconn says many will be doing traditional robot work: repetitive tasks like welding or spraying.  But many doesn’t mean all.  As Thomas Ricker notes over at Engadget, Foxconn will also deploy an untold number of FRIDA robots – small, agile devices that, purely coincidentally, can fit into work spaces currently occupied by humans.  “There can be only one logical conclusion to this endeavor,” writes Ricker: “FRIDA and its ilk will one day replace the millions of young Chinese workers assembling our gadgets.”

This may or may not be, but workers’ worries can’t be soothed by FRIDA’s YouTube videos, where the manufacturer gives ‘her’ a gender and says things like “FRIDA applies for a manufacturing position in your company.  She is a real team player.”  (None of this is helped by the fact that, in the video clips, FRIDA appears to be pleased to show off for her new employers.)

It’s stuff like this that may be fueling what appears to be a resurgent robo-fear. The American network G4TV features a periodic segment called Robot News Online, which promises “…all the news that will allow you to prepare for our eventual colonization and control by robots.”  The best-selling novel “Robopocalypse” is already in film pre-production, with no lesser than Steven Spielberg, a master of techno-angst, in the director’s chair.  And how could he not be, with tag-lines such as this:

“They are in your house. They are in your car. They are in the skies…Now, they’re coming for you.”

Bah-bah-bummmm!

Scary for some, but cooler minds return to the question: can computer brains and their robotic bodies do what humans do?  Everything humans do?  Sure, IBM’s Artificial Intelligence named “Watson” can help a doctor diagnose an illness, but can it comfort a troubled patient?  Do robots smile and help people find what they’re looking for and say ‘Please‘ and ‘Thank you‘?   Can a machine write news, or poetry, or make art?

Good Robots

You’re Not As Special As You Think

These are some of the questions Slate journalist Farhad Manjoo has been asking in his ongoing series of stories “Will Robots Steal Your Job?”  And his answers aren’t very comforting. “You are in peril,” he says at the outset, warning:

“At this moment, there’s someone training for your job. He may not be as smart as you are—in fact, he could be quite stupid—but what he lacks in intelligence he makes up for in drive, reliability, consistency, and price. He’s willing to work for longer hours, and he’s capable of doing better work, at a much lower wage. He doesn’t ask for health or retirement benefits, he doesn’t take sick days, and he doesn’t goof off when he’s on the clock.”

Anyone surprised that “he” is a robot may now leave the room.

In his series, Manjoo explores whether artificially intelligent machines (AIs) can do work like his father does (a pharmacist) or his wife (a pathologist).  Not surprisingly, it turns out there are still many things a computer cannot do.  For example, their huge databases mean they can diagnose an obscure disease a country doctor might miss.  But they can’t deliver that diagnosis with any empathy.  Computers aren’t funny.  Robots don’t gossip at the water cooler.  It’s these qualities, which AI innovator Ray Kurzweil calls “the subtleties and suppleness of human emotions and intelligence,” that make us special…at least for now.

However, both Kurzweil and Manjoo note that AIs are getting much better very fast…in fields as expert as medicine, and as human as nursing.  Still, for Manjoo, this isn’t enough.  And so, ever the writer – complete with the writer’s ego – he recently turned the tables on himself, asking if a machine can be a journalist:

“As a writer, I like to think of myself as having uniquely human skills. I write columns about stuff that human readers care about, and in doing so I try to elicit human emotions—joy, fascination, fury. Machines can’t yet mimic this sort of creativity. But as I surveyed efforts to automate journalism…I found that my job may not be beyond the capacities of a robot.”

Indeed, computers aren’t yet be up to the task of cracking open a scandal like Watergate, or conducting a daring interview, or writing that dazzling feature that people talk about for years.  But when it comes to the nuts and bolts of news – basic news like what, who, when, where and how – they’re surprisingly (or disturbingly) good.

Take, for instance, MotownBall – a baseball-crazy website with complete coverage of all the baseball games played by the Detroit Tigers.  “Tigers Top Yankees 5-3, Tie ALCS 1-1,” reads the headline of the top story, a summary of the most recent playoff game between the Tigers and the New York Yankees.  The first graphs read with zip and flair, as does the rest of the story:

“With the game scoreless in the top of the first, the Tigers broke it open when Cabrera rocked starter Freddy Garcia for a two-run shot, scoring Magglio Ordonez. Scherzer gets his first win of the postseason with a solid performance in which he surrendered no runs on two hits with five strikeouts and four walks.  Cabrera was terrific at the plate for the Tigers, going 3 for 4 with three RBIs and one run. Ordonez also had great output, going 3 for 3 with one run.”

Now consider that MotownBall is run entirely by computers.  At no point in the editorial chain – from story selection to headlines and text to web layout – does a living, breathing journalist contribute.  Its “journalists” are entirely digital.

Granted, the computers and their software were set up by real people; in this case, a company called Automated Insights, which runs one of these robo-web sites for every MLB baseball team.  Automated Insights engineers create the AIs – in their words, to “…transform raw data into compelling narrative content…” – but once they’re flipped on, the computers completely take over.  And they seem to deliver.  Sports journalism may be stat-heavy (a computer natural), but it also has a long tradition of zesty writing, which MotownBall has in abundance.  You could be forgiven for thinking someone with a soft spot for the Tigers wrote its stories.

Bad robot

The Man Machine

So if computers can write (MotownBall) and edit (see Google News), can one take my job today?  Not yet…thankfully.  Good reporting isn’t just compiling facts and deploying pungent verbs.  It’s talking to people – hearing what they’re saying and not saying in the stories they tell.  It’s using a bit of intuition to smell out a lie, or trying something unconventional to get at a truth.  There’s still a lot of human in a job where you try to understand other humans.

Which, in the end, may be the last genuine area where humans hold the edge over robots and AIs.  Despite any number of efforts, there still isn’t a computer that exists that can write poetry that’s worthwhile.   Or conduct therapy with someone experiencing clinical depression.  Or make a movie that anyone would actually pay to see…although sites like XtraNormal or Animoto demonstrate an unnerving skill in assembling punchy music videos with minimal input from us.  Robots don’t have emotions.  They never may (although they may learn how to mimic them very well), and without them, they’ll never understand the humans that created them.

So for the time being, my job is safe…at least, from robots.  And despite Farhad Manjoo’s concerns, so, too, are many of the other things humans do with their lives.  For all the dystopian worries, it’s unlikely Vonnegut’s vision of a world populated by working robots and unemployable humans will come to pass.

But it’s a certainty that as robots and AIs improve, we will have more of them around our workplace, and our homes.  Machines of super-human precision and endurance that will become better thinkers, problem solvers, and help-mates.  We will have to adapt to them, even as they learn how to interact with us.

As FRIDA’s manufacturers say of their dexterous new creation, “she is looking forward to cooperating with you.”

Let’s all hope it just stays at cooperation.