Work In Progress-Interview with Frank Levy of MIT
Wednesday, December 20, 2006Frank Levy is Daniel Rose Professor of Urban Economics in MIT's Department of Urban Studies and Planning. His research examines how computer technology and offshoring are reshaping opportunities in the labor market. He is co-author, with Richard Murnane, of The New Division of Labor: How Computers are Creating the Next Job Market. Darren Gersh interviewed Professor Levy for the NBR special, "Work In Progress." What follows is the transcript of their entire conversation; an edited version appeared in the broadcast report.
DARREN GERSH
Computers have been in the workplace for a long time. How do they change what
humans do compared to what machines do?
|
|
|
FRANK LEVY, MIT PROFESSOR
Dividing line is whether you can express the work in step-by-step rules, because
that's what computers do, they execute rules.
So if you look at issuing a boarding pass, you can break it down into a series
of steps first. Read the credit card number, then see if there is a match in
the data base. If the answer is no, then you bail out and send the person to
a desk agent.
There used to be a set of jobs where you could earn a good living just by following directions closely, and the change has been, if you could really describe a job in directions, step by step, that means you can program it or you can explain it to somebody else 10,000 miles away. So those are the pieces of work that are just disappearing.
DARREN GERSH
You had said what humans are going to end up doing is "expert thinking."
What is expert thinking?
FRANK LEVY
Expert thinking is how do you handle anything that can't be expressed in rules.
Suppose that a pair of grandparents come to the ticket counter and they go up
to a desk agent and they say we were supposed to meet our ten year old grandson
on a flight from Denver. We watched the people get off the plane, this kid did
not get off the plane, where's our grandson? Well, how do you write a computer
program to do that? There's no set of rules that you can find to program. The
desk agent will know a bunch of things through her experience, or his experience
to ask -- there may be a lounge that the kid is stowed in by a stewardess, maybe
that the kid wandered off someplace . . . but there is no way you could write
software to do that and on top of that, while the desk agent is solving the
problem, they also have to put the grandparents at ease and let them know that
she's on their side and she's going to solve the problem. If they looked at
a kiosk and the kiosk said I will find your grandchild, they would never believe
it in a million years.
DARREN GERSH
Computers have been around for a long time. And a lot of economists say the
people who do well with computers are earning more, and the people whose jobs
are routine and can be replaced by computers are earning less. Is the wage inequality
generated by computers going to keep growing or have they been around long enough
that maybe the wage inequality is about to abate?
FRANK LEVY
What you really have with wage inequality is a race between technology on the
one hand and education of the population on the other. If technology pulls ahead,
then you start seeing inequality grow. If education pulls ahead, then you see
it start shrinking. Right now, you would think it is going to keep rolling for
a while because technology is moving pretty fast and education is moving pretty
slowly.
But when you get up to inequality at the extreme upper tail, the top 1 or 2% that accounts for more than 16% of all income right now, that really is not a computer story. That’s something else altogether.
DARREN GERSH
There will be educated people in India, there will be educated people in Poland
and China. If you can ship the complicated problems over there, via the internet,
are we going to be competing with everybody everywhere?
FRANK LEVY
There are some kinds of expert problems you can send over a wire. and some kinds
of problems you can't. The problems that are most easily sent offshore are also
the problems that are easiest to computerize, because it is the breaking down
of a problem step by step that enables you to explain it to somebody else. If
it is a really complicated problem, you can never be sure that the person on
the other end of the wire really understands what you are talking about. So,
if we were to say write me a piece of software that can read a credit card and
match it up to a database, that is a well-enough specified problem so that somebody
10,000 miles away can understand it and write the software.
But suppose we say we've seen that customers are having a lot of problems using the airport check-in kiosk, make the interface more user friendly. What does that mean? To somebody sitting 10,000 miles away, how do they know exactly what it is to make something more user friendly. You have to come, you have to observe what people are doing. You have to be in their minds, in their heads to think about how they are approaching this and then you can begin to figure out what it is you have to change in order to do that.
DARREN GERSH
What is the role of face-to-face communication in an age of computers?
FRANK LEVY
Well, you have to begin with the idea that all the information you process,
is processed in context. If you're a doctor and your patient comes in to you
and it's not exactly clear what's going on, but the patient is somehow distressed.
If you look at the back and forth between the doctor and the patient, what you're
getting is the doctor is reading the body language and listening to the intonation
of the voice. Those kinds of interactions are setting a context to give you
a particular interpretation of information and that's very hard to do from a
distance.
DARREN GERSH
And so the jobs that will stay here are the ones that technology won't break
down into simple problems or the ones that require people to talk.
FRANK LEVY
Right, a lot of face-to-face, a lot of interpersonal communication, or the jobs
that can't be broken down into step-by-step processes.
DARREN GERSH
But, if the simplest problems are shipped over a wire, aren't those the entry
level jobs you need to have in order to move up to the complicated jobs? And
if those entry level jobs are gone, how do we develop expert thinking?
FRANK LEVY
That's a good question and I think what you're beginning to see is the economy
is beginning to understand that, that if we are really going to maintain talent,
we have to continue to grow our own and so you can't just sort of take the lower
end of stuff and send it offshore and then expect people to move up through
the system and be able to do the high end stuff, because this is very experience-based
learning that we're talking about.
