MENLO PARK, California--In a darkened room at AT&T Labs' facilities here, a sonorous voice intones a grave admonition.
"You have less than two ounces of milk in your refrigerator," it warns. "You may want to think about buying some more."
The voice being demonstrated is a synthetic one, reminiscent of the Machiavellian computer Hal in Stanley Kubrick's "2001: A Space Odyssey". The comparison is relevant in more ways than one: This technology represents the real-life voyage of AT&T Labs at the dawn of the millennium--an uncomfortable journey into unknown territory of its own.
A direct descendant of the institution that created radar, the transistor, lasers and hundreds of other innovations, AT&T Labs is for the first time selling the fruits of its research directly to the public. It might not win any Nobel Prizes, but the human-sounding voice software could help fill the Labs' coffers--and help justify its existence in a profit-driven 21st century.
The new business marks yet another crossroads for AT&T Labs, an institution once viewed as the world's premier research center, as it struggles to survive the economic storms once buffered by its corporate parent. The lab itself has been split and split again over the last two decades, with part of its staff and many patents going to Bellcore (now Telcordia Technologies) in 1984 and another to Lucent Technologies in 1996.
Its reputation for groundbreaking discoveries has diminished as financial and competitive pressures have put research on a more practical course. AT&T Labs is one of only a handful of world-class facilities linked to large corporations, such as IBM Research and Xerox PARC. Dozens of other research labs are affiliated with colleges, universities and the federal government but face their own financial obstacles, competing for public and corporate dollars as well as scientific prestige.
Now, as its parent company goes through a wrenching split of its own, AT&T's once-vaunted research division is entering a new round of experimentation with business strategies instead of science.
"Times have changed," says David Nagel, president of AT&T Labs and chief technical officer of AT&T. "We are a commercial entity."
AT&T Labs, and its predecessor Bell Labs, remain legendary among scientists and researchers. Like no other institution, its researchers churned out a century's worth of scientific and technological innovation.
Aside from the telephone, Alexander Graham Bell's invention, the labs that took Bell's name were responsible for inventing or doing early work on stereo sound, sound in movies, radar, lasers, the transistor, solar power, fiber-optic communications, Unix, the C++ computer language and countless other scientific firsts.
Many of the most lasting achievements had little to do with the narrow field of communications. Researchers Arno Penzias and Robert Wilson won their Nobel Prize for work that led to the theory of the Big Bang origin of the universe, for example.
That was another era, however.
"Nothing is like the old days," says Robert Lucky, corporate vice president of applied research at Telcordia, and a one-time executive director of a division of AT&T Bell Labs. "We never worried about making a profit. People's work was judged on the quality of their science. Those days are gone."
The world facing AT&T Labs today is a cold one. Parent AT&T is a chastened giant, in the midst of breaking itself four ways after its expensive move into cable television backfired. It is closely examining its businesses, trying to stem financial bleeding from declining profits in the long-distance phone market.
In its last quarterly earning reports, the company said that revenue from continuing operations had dropped 3 percent from the same quarter last year, to about US$13.3 billion. But its basic consumer phone business was far worse off. Revenue in that division dropped more than 23 percent compared to the previous year to about US$3.79 billion, the company said.
The company itself remains committed to the research division, although it does not release specific budget figures. It is still one of the largest corporate research facilities in the world, with 6,000 people on staff. By contrast, the vaunted Xerox PARC facility averages about 250.
Nor are corporate research dollars disappearing from the scene. In a report released late last year, the National Science Foundation (NSF) said that US research and development funding, adjusting for inflation, had risen 6 percent annually from US$176 billion in 1994 to an estimated US$249 billion. That was the largest six-year gain since statistics started being kept in 1953, the report said.
Nevertheless, AT&T Labs is increasingly being called on for its own contributions to the bottom line.
For the last several years, the company has been trying to align its research more closely with the communications business. Nagel, who came to AT&T from Apple Computer in 1996, has been instrumental in that goal.
"The majority of what we get is direct-development contracts with our business units," Nagel says. "If we can't agree on (how to proceed), we don't get funding."
That has led to a diminished profile for the research institution. Many outsiders largely see the researchers as problem solvers for AT&T's phone network, a kind of last-resort, long-term technical support staff.
"It is regarded as no more than a modestly tactical asset," said Howard Anderson, founder of The Yankee Group communications research firm and a partner in the associated YankeeTek venture fund. "The whole push over the last 10 years has been to solve tactical problems, not to get more Nobel Prizes."
Once Ma Bell divides itself into separate wireless, cable TV, consumer and business services divisions, the Labs will remain tied to the business branch. Part of the long series of internal divorce negotiations had focused around allowing use of lab technology among the divisions and which unit would own intellectual property rights. Nagel says these agreements have been written into the plans governing the breakup.
In the future, AT&T Labs can continue to serve AT&T Wireless or the cable TV division that appears to be teetering on the brink of a bidding war. But they will simply be one of the many to which the Labs licenses new technology, executives say.
"They will be like any other company, except we'll know them better," Nagel says.
The Labs still averages between one and two patents per day, an astonishing rate for any institution. And even as the institution has focused on practical communications technologies, its researchers remain respected across varied fields, from artificial intelligence to encryption and privacy.
The push now is to figure out how to turn that stream of intellectual property into cash.
Research labs have long licensed their discoveries out to other companies. CEO Mike Armstrong has been a big backer of this outside licensing--a standard practice in the software industry. Nagel cites a conversation three years ago, not long after Armstrong joined the company, in which the CEO urged him to boost licensing efforts.
In this type of relationship, AT&T allows an outside company to turn its technology into a product, in return for a license fee. SpeechWorks, for example, has licensed AT&T's earlier text-to-speech conversion technology for its own line of products.
AT&T Labs, like other research institutions, also is open to spinning new companies out of technologies that don't quite fit into the parent's core business. One such technology, a broadband Internet-based phone system that would give simple Web-like features to a telephone with a small screen, is being readied for this kind of treatment, Nagel says.
But the real innovation, at least for the Bell Labs descendant, is the idea of setting up a division that actually makes products out of its research, and sells them to the public, an idea that borders on heresy for the old Labs philosophy.
Nagel's group made its foray into this terrain last week, with the release of the Natural Voices text-to-speech technology.
Only a tiny fraction of the Labs' budget, about three or four positions, will be dedicated to sales and marketing. The product will likely be sold more effectively by third parties, the organizations' president conceded in his demonstration to members of the press. But it also serves as a model for the facility's new role.
Others are watching the approach in the industry as well. AT&T isn't the only giant with financial difficulties and a giant research arm. The future of Xerox PARC, the facility that gave the world the graphical user interface adopted by Apple Computer and Microsoft, has been clouded as its parent company grapples with financial losses, for example.
AT&T Labs is in the middle of "a really interesting experiment", said Johan de Kleer, a principal scientist at Xerox PARC. "Just like we're experimenting (with science), we're all now experimenting with new ways to make money."
Whatever their success, critics say these survival tactics are masking a more serious problem. AT&T Labs and its ilk were once one of the primary sources of research driving technological progress in ways far beyond what corporate strategy-makers might dream possible. That freedom is all but gone, say research veterans, and that bodes ill for science as a whole.
"The ability to roam the scientific pastures is endangered by the whole business process," Telcordia's Lucky says. "(AT&T Labs) still has good people. They do good work. But they're an endangered species."