When Big Science Fails To Deliver, Researchers Invent New Strategies

Harold Furth could see the hard times coming. Oil prices were dropping and, with them, the prospects for funding research into alternative energies such as fusion, his specialty. So, like others in his field during the early 1980s, the head of the Princeton Plasma Laboratory started promoting his program as good basic research, rather than as the pursuit of a working fusion reactor. Meanwhile, researchers in pursuit of a malaria vaccine had their own troubles. Serious scientific obstacles had emerged, and the field was dissolving in chaos. Faced with Draconian funding cuts in recent years, policymakers like biochemist Charles Chambers at the American Institute of Biological Sciences (AIBS) responded in a manner similar to Furth’s: Rather than emphasizing the prospects of a working vaccine in the foreseeable future, they began touting the value of their malaria studies to basic immunological research. And throughout the last two decades, intelligence pioneer John McCarthy watched his own discipline evolve from an ambitious quest for reasoning machines toward a series of short-term demonstration projects, to appease a Defense Department impatient for quick results. He himself held out for long-term goals, but paid the price in diminished funding. These three scientists have experienced firsthand the phenomenon of too-slow science—science that doesn’t deliver quickly enough its promise to revolutionize its discipline. The fields of fusion, malaria, and artificial intelligence research have, like many others, been characterized by periods of overoptimistic claims and subsequent disappointing results. As confidence in the science eroded at funding agencies and on Capitol Hill, some researchers have been forced to redirect their research and rewrite grant proposals. Indeed, in the case of malaria research, failed, overblown expectations have subjected some of its leading figures to criminal investigations. These frustrations have forced scientists to develop new strategies—with new promises—to ensure the government funding that will keep them in business. A Squeeze On Fusion “I call it the ‘manic-depressive cycle,’ “ says Furth, whose Princeton, N.J., laboratory is the nation’s leading fusion facility. Since the early 1950s, the field has weathered a roller-coaster ride of shifting priorities and varying experimental success. When things were good, they were very good: In the late 1970s the Department of Energy was spending nearly a half-billion dollars a year on fusion, compared with the $330 million planned for 1990. But Furth is now battling to maintain level funding for his lab. His top priority is the oft-delayed next step in fusion: construction of the $400 million Compact Ignition Tokarnak (CIT). Such a machine, say fusion proponents, would sometime in the next century achieve a point at which more energy is generated than is used. That break-even point is the grail that has eluded the program for more than 40 years. “During the ‘manic’ phase, you imagine that the budget will rise forever,” Furth says. And in such an atmosphere, he says, a fusion scientist’s thoughts turn to huge and expensive demonstration projects. For the last decade, however, researchers have been mired in the other side of the field’s split personality. Behind the decline are the Three Horsemen of fusion’s apocalypse: prices, privatization, and promises. Oil prices have plummeted, stalling the search for alternative sources of energy. The Reagan-era belief that industry should pay to develop products with commercial value has gutted the DOE research program. And the fact that the program has repeatedly fallen short of its goal—for both scientific and funding reasons—has undermined support in Congress. Although the research is progressing steadily, overblown promises made early in the program still haunt the fusion scientists. One 1958 report, for example, anticipated a full-scale power-producing thermonuclear device by 1978, given little more than “hard work and a sprinkling of good luck.” Without dramatic results to tout, scientists have learned to swim in the currents of federal funding. “We tried to see what it took to interest the politicians,” says physicist Stephen Dean, director of Fusion Energy Associates in Gaithersburg, Md., and a former head of DOE’s magnetic fusion program. “For [former President Jimmy] Carter, it was energy development. For [former President Ronald] Reagan it was good basic science.” Because Reagan believed that industry—and not the federal government—should fund energy development, says Dean, “we sold [fusion] as basic research. We said it was too long-term and expensive for industry.” The new approach called for an emphasis on investigations into the basic physics of confining superhot plasmas, rather than the engineering of large production machines. But although everyone agrees that there are many important physics questions to be answered in fusion, some worry that basic research will become a “cerebral pursuit” that will delay them from developing fusion as a practical energy source. Robert Hirsch, who directed the DOE program through much of the 1970s, is among those who have fought the dominance of basic research. “My motivation for every moment was a practical power system. I never changed my tune,” he asserts. Hirsch, now a vice president at the Atlantic Richfield Company, pushed for a major demonstration project that would make both symbolic and technical strides toward the break-even point. Most impor-tantly, Hirsch hoped it would force researchers out of the abstractions of plasma physics. Indeed, fusion’s fortunes soared under Hirsch and his successor, Edward Kintner. Both physicists were aided by skyrocketing oil prices that drew attention to fusion-derived energy. As a result, by the late 1970s the program had grown fivefold. But when oil prices declined over the following decade, so did funding for fusion. Faced with dim prospects for new machines, scientists have now found it politic to stress the value of basic research. “The strategy is to say that ‘the physics isn’t there, so I’m going to work on pure science for a while,’” says Kintner, now executive vice president of the nuclear divisionof General Public Utilities. “But [the researchers] don’t realize that you can’t justify that kind of spending on science alone.” Malaise Over Malaria If losing one’s funding is depressing, watching your colleagues being indicted is worse. Since 1966 the Agency for International Development (AID)—the government’s chief source of funding—has spent nearly $100 million in a search for a malaria vaccine. Desperate to justify the money spent, AID held a 1985 press conference to announce that two New York University researchers had found the possible key to a genetic immunity to the disease in humans. “It is not unreasonable to expect clinical trials to begin within the next two years,” said the press release, “and to have a vaccine ready for widespread use by 1989.” Malaria scientists now wince when they recall those claims. “There are a lot of immunologists who will tell you that there will never be a malaria vaccine,” says Chambers, executive director of AIBS, which coordinates the AID program. “I, myself, am very skeptical that a malaria vaccine, in the same sense as a vaccine for polio, will ever be found.” The breakthrough promised in the 1985 press conference has yet to materialize. But a lot—most of it bad—has happened since then. The program has been rocked by charges of fraud and mismanagement (The Scientist, July 10, 1989, page 1). Two researchers and a program official are under indictment, and a recent investigation by the federal General Accounting Office recommended that AID refrain from making any longterm funding commitments to malaria vaccine research until the . National Academy of Sciences finishes a two-year review of the program; From a high of $13.5 million in 1985, the program’s budget has fallen to $9 million, with further cuts anticipated. Like their colleagues in fusion research, malaria scientists have responded to this threat by circling the wagons and turning to basic science. Rather than offering deadlines or specific goals, researchers now talk about the value of basic immunological science. “I think we learned our lesson,” says Chambers. “Now, when asked [for a status report], we say progress is being made.’ We’re not on the verge of a solution. All the vaccines that could have been found by accident have been tiled. There have been hundreds of things stuck in people’s arms, but they didn’t work.” To restore flagging federal interest in the field, malaria scientists are even hopping on the latest scientific bandwagon: the possibility of global warming via the greenhouse effect. If increased levels of carbon dioxide in the atmosphere actually boost temperatures worldwide, the argument goes, tropical mosquitoes might inyade North America. “If the Tropic of Cancer moves up to Texas, you’ll have malaria there,” warns one researcher. Others have broadened their approach to include the control of mosquitoes and the use of drugs and prophylactic measures to help affected populations cope better with the symptoms of malaria or avoid the disease altogether. Years In The AT Wilderness Compared with the search for a malaria vaccine or a sustainable fusion reaction, the often-maligned field of artificial intelligence AI is a veritable success story. After a 30year history that juxtaposed the rosy popular picture of intelligent machines with a string of experimental failures in speech recognition, robotics, and “expert systems,” Al is finally gaining acceptance in the marketplace and recapturing the respect of computer researchers. But a generation of meager re suits has taken its toll. Dozens of AI startup companies lost millions of dollars and sank into bankruptcy as the technology of speech and language recognition proved more difficult than anyone anticipated. In 1987 alone, nearly a dozen AI companies reported losses, and layoffs approached 20% in the expert systems industry. Meanwhile, Wall Street lost confidence in the field, and corporate funding dried up. Belief in AI “is a religious issue rather than a technical or political one: Wall Street, however, is short-term profit driven,” says AI pioneer Marvin Minsky of the Massachusetts Institute of Technology. Al research was even the unwitting victim of congressional displeasure with the classified work ihat the Pentagon funded on college campuses during the Vietnam War. The 1971 Mansfield Amendment, intended to restrict Defense Department funding to projects with demonstrable military applications, had the additional effect of forcing AL scientists to shift frombasic research to short-term projects with clear and achievable goals. The impact was greatest on the Defense Advanced Research Projects Agency (DARPA), which has traditionally been the chief source of federal funding of artificial intelligence. Because of DARPA’s change in emphasis to short-term research, we are not as far along 18 years later as we might otherwise have been,” says Carnegie-Mellon University computer scientist Tom Mitchell. At Mitchell’s own institution, for example, researchers have recently shown considerable success with a DARPA-funded “autonomous vehicle.” It’s a van that can slowly drive itself down tree-lined streets. But although the project drew applause for its unprecedented achievements, some question how much it has advanced the field. The problem, according to Mitchell, is that “while DARPA intended to push forward the research on computer vision, what they got was some very specialized algorithms for finding the edges of roads.” Stanford University AI pioneer John McCarthy was among the first to regret over-optimistic promises. He recalls one early $1.5 million proposal that he submitted to DARPA to design robots that would be capable of assembling electronics kits. He got the grant, but autonomous robots turned out to be far more difficult than he had expected. Even today, more than 20 years later, robots still can’t be entirely trusted with a soldering iron and a screwdriver. That experience has taught McCarthy to resist the temptation to propose short-term applied projects and, instead, continue his basic research on common-sense logic. But his decision comes at a price. “My own work has not changed its orientation much [to reflect DARPA’s priorities], so I get a lot less funding,” he says. Mitchell has tried to find a middle ground. “My way of surviving in the current funding climate is to find a project with a clear consumer product,” he explains. In his case, that means work on an automated vehicle for a future NASA mission to Mars and research focused on robots that can be useful in biologic and chemical laboratories. No matter what type of research they pursue, AL scientists have learned the importance of avoiding the rhetoric that overheated the field in the 1960s. “Al is grappling with some fundamental problems,” says Jack Schwartz, who headed DARPA’s Al program before leaving for New York University earlier this year. “Creating intelligence,” he says, “is like creating life. It’s possible, but it takes time.”

When Big Science Fails To Deliver, Researchers Invent New Strategies

Interested in reading more?

Become a Member of