In part I of this two-part blog series, we highlighted some modern-day examples of intelligent automation, one of the five technology trends highlighted in our 2016 Accenture Technology Vision. At first blush, they seemed like perfect examples of machines taking jobs from humans; but in fact, they demonstrated the ways that intelligent automation can open up opportunities for humans to use their skills and experience more effectively. Rather than robots taking jobs from humans, humans and computers were being productive together.
Still, fears remain that computers will ultimately steal our jobs—that despite past instances of automation being used together with humans, we’re on the cusp of wholesale replacement. Some of this is no doubt due to the speed with which robotics and automation are growing: One forecast projects the US market for Artificial Intelligence to grow by 53 percent by 2020.iTesla has already released software updates that allow their cars to largely drive themselves, and Elon Musk predicted in December that the company will achieve total autonomy in two years. Google has already introduced a self-driving car without a steering wheel or pedals, and the National Highway Traffic Safety Administration has determined that a computer system can qualify as the legal driver of a car.
Despite past demonstrations of robots not displacing humans in the economy, this level of growth stokes a common fear: that robots and automation will result in wholesale job elimination – that the only reason a company would deploy automation is to cut costs by cutting people. In other words, despite all past evidence to the contrary: the robots are coming, and this time it’s different.
Stop me if any of this sounds familiar: machines are on a path to “replace humans”; “robot evolution” will bring about “large-scale unemployment”; or the catchy argument that we’re “innovating ourselves out of relevance.” These types of arguments have become a common trope, with even Stephen Hawking and Elon Musk chiming in about the risks of AI. No doubt these fears are hastened by impressive technological advances and the speed with which we learn about those advances today.
Fortunately for the human race, the rumors of our demise are greatly exaggerated. While it is reasonable and valid to be concerned about job loss as a result of automation, intelligent automation offers substantial benefit in the workplace—and not by shoving humans aside and plunking a computer in their place. Let’s take a look at some of the most common arguments in this space, and see how they hold up to scrutiny.
Technological Unemployment: an argument almost as old as machines themselves. This is the argument that boils down to “Look what happened to the horse” – that is, when automobiles became commonplace, the horse was marginalized from relevance as farm equipment, cars, and trucks could do everything that horses had been used to do. Now, some studies see as much as 47 percent of jobs in the US economy being displaced by technology in the near future.ii
Let’s start with the obvious here: people aren’t horses. Humans can adapt and re-orient for new challenges. More importantly, in the historical context of a “workforce,” horses were a tool that humans used to accomplish their own tasks: Plowing a field, transporting goods, or transporting other humans. We can’t compare a tool with the people wielding it. Using the horse example to suggest that robots will run humans out of the workplace is like saying that humans will become irrelevant as evidenced by the power drill replacing the screwdriver. Clearly, that did not happen; we developed a new tool that let us do things faster, but we’re still here using it. And automation is also a tool of our own creation—one that we’re designing against our own skills. Which brings us to:
Thinking Robots, or: “Watson beat us at Jeopardy!” Yes, IBM’s Watson thoroughly trounced two Jeopardy champions, and Ken Jennings (who finished in second place) even joked about welcoming “our new computer overlords” in his Final Jeopardy answer. But this is a situation where the final result doesn’t tell the real story.
We’ve been using games to benchmark what machines can do for decades (think back to Deep Blue playing chess, all the way up to Google’s Deep Mind beating the human “Go” champion this year). This gives us what seems like a helpful metric for gauging computers’ ability to “think” and solve problems—so when Watson won Jeopardy!, there was an immediate flurry of “Watson is going to put us all out of a job” articles.
In fact, using performance at games as a proxy for beating us out in the economy is incredibly misleading, and here’s why: We create games for us to be bad at. Humans like challenging games, and we design them specifically to be difficult for us to win. Kasparov lost to Deep Blue after people “taught” it the chess rules and possible sequences of moves. Machines can extrapolate and hold all possible moves in memory. People can’t do that; it’s exactly what makes the game hard, and fun.
Similarly, Watson winning in Jeopardy! was a reflection of two things that computers can beat humans at consistently: Recall (Watson was fed 200 million pages of content that took up 4TB of disk space), and speed. Once its human designers taught Watson how to play Jeopardy, it’s not surprising that it won. And keep in mind that Watson wasn’t really responding to Alex Trebek—it got the Jeopardy prompts as text files. Still impressive, but the system was using its 2,800 processor cores to generate responses to flat-text inputs, not naturally respond to human vocal cues.
So yes, computers have now beaten us at chess, Jeopardy!, and Go—all games designed to be difficult for humans. Here’s the flip side: There’s an incredibly simple game that humans trounce computers at time and time again. We don’t play it, because it’s easy for us and incredibly boring—identifying an object in a photo. A six-year-old can look at a photo and tell you what’s in it. Not so for computers—they usually fail, sometimes to hilarious effect.
Of course, none of these examples can completely allay fears about automation. And it’s both fair and reasonable to keep those concerns in mind while deploying robots and intelligent systems across an enterprise. But the biggest winners will be those who deploy this technology to get the best from both computers and humans, to the benefit of everyone. That’s intelligent automation.
Ari Bernstein is a co-author of the Accenture Technology Vision 2016. See the Accenture Technology Vision 2016 here.