HIAWATHA BRAY UPGRADE;
IN DEFENSE OF DARPA PROGRAMS

BY HIAWATHA BRAY

The Boston Globe
BUSINESS; Pg. C3
August 4, 2003, Monday ,THIRD EDITION


In 1970, the Defense Advanced Research Projects Agency, or DARPA, went about its
business in placid obscurity, which is just as well. It meant that Americans
weren't regaled with inaccurate scare stories about various DARPA projects, some
of which turned out rather well. That global computer network, for instance,
which we now call the Internet, or the invention of military airplanes invisible
to radar.

But in the past half-year, DARPA has managed to make the front pages twice, with
lamentable results. Last week, the agency backed off a startling plan to set up
a kind of futures market that would allow investors to earn profits by betting
on the likelihood of terrorist acts. Earlier this year, Congress moved to scale
back another DARPA program to scan commercial databases filled with information
on millions of Americans, in an effort to spot patterns of terrorist activity.
Between them, these two programs aroused so much political outrage that the man
heading them up, former National Security Adviser John Poindexter, has been
forced to resign.

   Good news for American liberty, say DARPA's critics. Maybe. But it's
certainly a defeat for the kind of daring, edge-of-the-precipice attitude that
is supposed to motivate the researchers at DARPA. The agency finds itself
condemned for doing its job -and not too badly, either. For both of these
controversial proposals have considerable merit.

Consider the database-scouring plan, originally called the Total Information
Awareness program. Press accounts falsely described it as a proposal to create
dossiers on millions of Americans. Nonsense. DARPA wanted to use information
gleaned from past incidents of terrorism to identify patterns of financial or
travel activity that might provide tip-offs. Then they wanted to see whether
powerful computers could spot the same patterns buried in vast amounts of
commercial data.

DARPA wanted to use made-up data of its own creation, to see if such a system
could work. It might not: Many computer experts are skeptical. But what was the
harm in trying? If the system proved practical, there'd have been plenty of time
for Congress to set limits on its use. Instead, DARPA was condemned merely for
proposing the concept.

The terrorism futures idea sounds even more shocking, but it turns out to be
more firmly grounded in solid science. It's based in part on the work of Nobel
Prize-winning economist Vernon Smith of George Mason University, who thinks that
prices on futures markets embody remarkably accurate information about the real
world. People with a financial incentive to get the facts right will constantly
ferret out new information and use this knowledge to get on the right side of
the trade, thus bidding prices up or down. So a market made up of well-informed
political and military experts - maybe even a few terrorists - would be likely
to provide accurate warnings of future events.

Of course, commodities traders have long used markets to protect themselves from
future risks. But the idea is being applied ever more broadly by academic and
commercial institutions. The Iowa Electronic Markets at the University of Iowa
has accurately predicted the outcomes of the past four presidential elections,
based on the prices of futures contracts associated with each candidate.

And a Boston company, Incentive Markets Inc., has begun selling a futures market
program for use inside corporations. Say managers want to know if a complex
project will be finished on time and on budget. Employees are assigned a certain
amount of play money, called "marks," which can be redeemed for real-world
goodies like days off. They're then allowed to buy and sell futures contracts on
the corporate computer network, betting either that the job will be done on
time, or it won't be.

An employee who thinks things are going well will invest in a "yes" future, but
one who knows about an unexpected glitch will bet "no." The relative prices of
the two futures will give managers a good idea of what to expect. In the
collective quest for profit, the truth emerges. Why not the truth about
terrorism?

DARPA has been knocked out of the terrorism futures game, but there's nothing to
stop some businessperson from launching a similar project. To attract lots of
investors, he'd probably promote it boldly and without shame. An entrepreneur
can afford to do things like that. Not so the zany innovators at DARPA, who've
no doubt learned a new appreciation for stealth technology.

Hiawatha Bray can be reached at bray@globe.com.

GRAPHIC: PHOTO, Senators Byron Dorgan (D-N.D.), left, and Ron Wyden (D-Ore.) at
a news conference last week sought to end a program aimed at predicting possible
terrorist events. / AP PHOTO