Is technology worsening America's jobs crisis?
Yahoo News spoke Thursday to McAfee (right) -- a top research scientist at the Center for Digital Business in the MIT Sloan School of Management -- about why a busboy might have less anxiety about his job security than a lawyer should, how we can make the automation trend work for us, not against us, and why this all this is good news for Lady Gaga.
YN: What types of jobs are currently being affected most by the development of new technology?
AM: The classic examples have to do with customer interfaces. So, clerks, cashiers, salespeople. We can now buy stuff online without talking to a person or interacting, check into our flights at the airport, check out of Home Depot, and just do a lot of these kind of routine interactions without having to interact with a flesh-and-blood human being.
Computers are just improving by leaps and bounds in their ability to do that stuff. And one of our points in the book is, if you think
Watson and the
Google Car and
Siri are impressive, you ain't seen nothing yet.
[These jobs] are a big source of employment. Retail's a huge industry in the United States. They are good old-fashioned middle-class jobs in America. And we think that this amazing encroachment of technology is part of what's making the middle class feel justifiably kind of precarious and nervous.
In addition, we're starting to see that, already, some of the pattern recognition technologies that we talk about in the book are having an impact in a field like law. So
e-discovery is becoming a big deal. And anecdotally, we're hearing that a lot of companies and law firms are using software instead of rooms full of lawyers to look over documents as part of a discovery process. And we include a quote in the book from a guy who is now using e-discovery who went back and applied it to a huge set of documents that he had previously used human lawyers for. And his conclusions was that the humans were only about 60 percent accurate, and a huge amount more expensive. So he said, 'I spent a lot of money to do a little bit better than a coin flip.'
YN: This is hardly the first time in history that there's been concern about machines replacing humans on the job. In the past, those concerns have generally proved to be unfounded. Why might this time be different?
AM: . The question, what's different now, is a critical question. Even though there was a huge amount of automation historically, when you look at the total list of skills that you might hire a human worker for, all this automation barely encroached on those skills, and especially the mental or the cognitive ones. What's interesting to us, and the reason we wrote the book, is that suddenly we start to see digital technologies encroaching on skills where they never played [a role] before, where humans alone had these skills. So we talk in the book about things like more complex communication: understanding human speech, responding in human speech, translating human languages. Up until pretty recently, computers were hopeless at that, and now we're in an age of Siri, and all kinds of very powerful translation tools. They're not perfect--none of these is perfect--but they're pretty good.
Another skill we talk about is pattern recognition. Historically, computers have been terrible at, for example, looking at a large body of documents and finding common threads in them. Or finding whether or not there was a pattern of deceit or malpractice. That's exactly what e-discovery software does. And we look at Watson, this Jeopardy-playing supercomputer. It has this astonishing ability to sit on hundreds of millions of documents and extract patterns and meaning from them.
So there's this large-scale, recent rapid encroachment into stuff that computers have never been good at before, and where humans were the only game in town. And when we look at what a lot of the middle class, even more educated white-collar workers do, we see them doing complex communication and patterns.
YN: You argue in the book that in addition to reducing job growth, automation is likely to increase inequality. Why is that?
AM: We talk about three different ways that technology can increase inequality:
It favors more highly skilled workers over less skilled workers. So if you have an MBA, a PhD in computer science, a lot of these advanced degrees, especially in
STEM fields, computers are great. You use them to do your work, and actually your salary has gone up in recent years. For mid-skill workers, for this large middle class in this country, when you look at them, they are doing stuff that, again, computers are encroaching on. That tends to drive down the wages that employers are willing to pay them, and tends to make them more likely to wind up unemployed. The fact that technology typically favors more highly skilled workers--this is one of the things that's going on.
Technology also rewards superstars, and we mean that in a couple different ways. It rewards Lady Gaga and Yo-Yo Ma, because they can suddenly replicate their work and sell it to millions and millions of people. And so if you like Lady Gaga a little bit more than the second best, or Yo-Yo Ma a little bit more than the second best, you're not gonna buy that much of the second best because you always have access to Yo-Yo Ma.
And then finally, when we look at financial services, it's impossible to be a modern investment banker or trader and not have a lot of technology at your disposal. You can't do CDOs and CDSs and all these esoteric financial products without a lot of technology whizzing away in the background. Now we can talk about whether those were good things or bad things--and I think it's pretty clear a lot of those were actually bad things for the economy. Our point is they did lead to superstar effects, and they were supported by technology.
AM:
David Autor has done the best work in teasing apart what's going on. So it's not just that high skill wins out and low skill loses. It's like you say, actually, if you're at the very lowest levels of education and skill, you're not as bad off, or things are not getting as bad as quickly, as if you're in this big middle distribution. And our explanation is that if you're a dog walker or a restaurant busboy, classically it's a low-skill, low-education job, but it's not one that technology's going to displace. So, for now, those jobs look relatively safe. And when we look at home health aides and food service workers and a lot of these employers that are still employing a lot of people, we think it's at least partly because technology's not available to automate those kinds of jobs away.
YN: Clearly, we're not going back to a world without technology. And technology brings a lot of benefits, in terms of lower costs, improved quality of life, etc. So how we do think about how to balance those upsides with the downside of its impact on jobs?
AM: I want to underscore what you just said. Our point in the book--and we want to stress it over and over again--is that technology is not bad. Technology is not the culprit here. Technology is growing the economic pie and it's improving our standard of living. We think that's fantastic. The last thing we're advocating, is, stop the innovation, shut off the machines, or do anything like that.
You bring up this central question, what do we do about this, given that the average worker is getting left behind by the cutting edge technologies?
We put a bunch of recommendations in the book for policy changes that basically help people acquire the skills to be good workers in this technology-intensive age. The kinds of skills you need to work in a factory these days are not what you needed 30 years ago or 50 year ago. One of the amazing things I keep hearing is people who want to start factories in the United States--these are really highly automated, very productive factories, they only need a couple hundred or maybe at most a thousand workers. They can't find those workers, because they can't find people with the right skills. So we absolutely need to shift education and shift the skills we're imparting to both young people and adults out there in workforce already.
We also do hear over and over again from executives and business owners that there is this kind of thicket of regulation and red tape that you have to go through if you want to start something up and employ some people. We need to work very hard to clear out that thicket.
This is a also great time to invest really heavily in infrastructure, A) because we need it, and B) because computers are still pretty lousy at repairing roads and bridges and putting sidewalks in. That sounds like we're rabid Keynesians or something, but we think that's a great idea.
YN: So what does all this mean for the kind of country we're likely to have in the future?
AM: As long as America's been around, there's been this social contract, where if you are willing to work, there will be a job available to you. Americans don't think the deck is stacked against them. We believe that we're the land of opportunity. That's fantastic.
But I believe that we're heading into the next chapter of our economic history, where for a lot of people who don't have exactly the right skills or have been left behind in this race against the machine, there might not be a job waiting for you, at least in the classic sense that we're used to thinking about a job. And we had better start thinking long and hard about how we react to that as a society and an economy.
This interview has been edited and condensed from the original transcript.
