A Computer Scientist Uses His Art to Question the Embrace of Technology
Ken Goldberg's research is all about facts, but his online works challenge viewers' reality
By LAWRENCE BIEMILLER
Berkeley, Calif.
Back away from the big painting in Ken Goldberg's office and its heavy black lines resolve into a man wearing a homburg and smoking a cigar. It's a portrait painted partly by a robot Mr. Goldberg created in 1992, and the subject is William Mulholland, the man who built a 238-mile aqueduct to bring water to Los Angeles from the Owens Valley.
But that's not the Mulholland story Mr. Goldberg is eager to tell a visitor to the University of California's campus here. Mulholland was also the man who designed the St. Francis Dam, whose 1928 collapse Mr. Goldberg describes as "the worst man-made disaster in California history." More than 500 people were killed when a wall of water swept across Ventura County in the middle of the night.
Today the dam's ruins are overgrown and nearly forgotten, and the cause of the collapse will probably never be known. Mulholland was never formally accused of wrongdoing, but even so, "he surveys the damage, resigns, and never speaks to anyone again," as Mr. Goldberg puts it. He stares up at the portrait for a moment as he prepares to leave for his 11 o'clock industrial-engineering class, in which he will explain to his production-methods undergraduates how to calculate tolerances in manufacturing processes.
Back away from Mr. Goldberg himself, and it's harder to know what you're seeing. Are you looking at a robotics researcher whose 1994 "Mercury Project" is believed to have been the first robot controlled over the Internet? An industrial engineer who holds three U.S. patents? An online artist whose humorous "Ouija 2000" will be included next week when New York's Whitney Museum of American Art opens its high-profile 2000 Biennial Exhibition?
Mr. Goldberg, it seems, is intent on being all those people and then some. A 38-year-old associate professor of industrial engineering and operations research, he's a man with two complete vitae -- the longer in computer science and engineering, the shorter in art -- and more projects and interests than most people could hope to keep track of. Curly-haired and inclined to dress somewhat stylishly, he wears a ring encircled by the words "EAT SLEEP CREATE."
As a professor, he will spend an hour this afternoon discussing parts feeders for assembly lines with students in a graduate-level robotics class. As an artist, he's well known for a series of online works -- most created with the help of his graduate students -- that are intended to be skeptical of technology and of popular expectations for it (http://www.ken.goldberg.net). As a sometime philosopher whose enthusiasm for ideas can be contagious, he has edited a new book of essays called The Robot in the Garden: Telerobotics and Telepistemology in the Age of the Internet, which is due out this month from MIT Press. The book, he says, considers epistemology in new ways, asking: "What can you know when the knowing is at a distance, especially when the knowing is mediated by the distance?"
Such questions make Mr. Goldberg's face light up. The "Mercury Project" -- his first online artwork, created while he was still an assistant professor of computer science at the University of Southern California -- allowed visitors to control a robotic arm in a sand-filled container, using blasts of compressed air to dig for artifacts that were described in a fictional "backstory" involving a prominent paleohydrologist, a former theater major, and a nuclear test site in Nevada. Almost as soon as the project was connected to the Internet, however, users surprised Mr. Goldberg by asking repeatedly whether what they were seeing on their monitors was real.
"What I suspect is going on," one user wrote, "is that you have two large images, one of the dirty site and one of the clean site. You only transmit the portion of the image near where the user clicked, and you alternately return portions of the clean image if the user has blown air in the vicinity." Although the user later satisfied himself that he was actually operating the robot arm, other doubters followed.
After the "Mercury Project," Mr. Goldberg and his U.S.C. graduate students produced the "Telegarden," which set a robotic arm in the middle of a small plot of earth. "From anywhere in the world, one can access it to plant seeds and water them," says text on the site, which is still online (http://telegarden.aec.at).
"I always intended it as ironic," Mr. Goldberg says of the "Telegarden," which was widely publicized. "The last thing you wanted to do over the Internet was garden. I was thinking, This is dystopia. But the irony, I think, was missed by most of the people who ever played with it."
And again, he says, users wrote to ask if what they were seeing was real. "What I found was that when people were most skeptical, they were most involved in the garden, watering a leaf and going back to see if the watering was consistent -- that's when I started to play with skepticism."
Not everyone plays with skepticism the way Mr. Goldberg did, however. In 1996, he created a model of Fallingwater, a house designed by Frank Lloyd Wright, that is no bigger than the width of a human hair -- a work that viewers could see only with the help of an electron microscope. The project, called the "Invisible Cantilever," raised a new kind of telepistemological question, Mr. Goldberg says -- a question not of distance in space, but of scale. If something is too small to see with the eye, do you know you're really seeing it?
Even more challenging is a 1998 work called "Dislocation of Intimacy," which is also still online (http://www.dislocation.net). Physically, the piece now occupies a corner of Mr. Goldberg's lab in the basement of Etcheverry Hall here, but those who have seen it properly displayed in galleries have confronted a large, sealed box with an "umbilical cord" running to a wall. In what Mr. Goldberg refers to as an inversion of the normal gallery-going experience, the only way viewers can fully experience the piece is by taking home the address of the work's Web site. Through it, anyone can turn five lights in the box on and off and see blurry images of objects inside.
Or can they? "It could be so easily faked," Mr. Goldberg volunteers. And is it? "I'm not prepared to say," he replies. "If you put up a site where people don't care whether it's real, you haven't accomplished very much."
Such reticence, he admits, comes hard to someone with an engineering background. "As an educator and an engineer, I'm doubly responsible for making things clear. If you're reading a paper, you nail the author for being unclear. But the art world is about a different mode of presentation -- it's not about explaining.
"The language of art is essentially silence," he says. "As an artist, the hardest thing has been learning how to say less and let a work be self-standing. People want to know, What were you thinking? Why? But often a work means less after an artist has spoken." Talking about a piece, he says, can close off alternatives. "It's not really up to the artist to say, 'This interpretation is right. This is wrong.'"
The more ambiguity Mr. Goldberg's online art creates, the more attention it seems to draw. In "Ouija 2000," which was commissioned by the university's art museum, he offers an online Ouija board with which multiple users can interact simultaneously. The Ouija software asks questions and then averages the mouse movements of the various players to direct a robotic arm -- also in Mr. Goldberg's lab here -- with a standard Ouija planchette attached.
The questions are as entertaining as Mr. Goldberg and his students could make them without letting players ask questions themselves -- Mr. Goldberg had no way of screening out off-color submissions -- and the software is set up to confirm the answers that players see the planchette giving on their monitors: "Will there be a devastating computer virus in 2000?" "The spirit says no." "Will a luxury cruise ship sink in 2000?" "The spirit says no."
To the layperson, such questions might just seem like good fun. But Heidi Zuckerman Jacobson, curator of the museum's experimental Matrix series, takes "Ouija 2000" more seriously. "Rather than asking his viewer/user to trust him and what he presents (to suspend disbelief)," Ms. Jacobson says in her curator's statement, "Goldberg encourages, perhaps even taunts, them into a resumption of disbelief to counteract the overwhelmingly enthusiastic and uncritical embrace of technology." The piece, she adds, "is intended as a critique of conventional notions of contemporary spirituality, technology, and the corporate application of science."
"Ouija 2000" is being taken seriously outside the art world, too. Besides winning a place in the Whitney Biennial, the work is the topic of a paper called "Collaborative Teleoperation via the Internet" that Mr. Goldberg will read next month at the Institute of Electrical and Electronics Engineers' International Conference on Robotics and Automation.
While the Ouija site itself cautions that "many people feel that Ouija boards summon powerful forces -- you have been warned," the powerful forces discussed in Mr. Goldberg's paper are chiefly mathematical. "We describe a system," the abstract says, "that allows a distributed group of users to simultaneously teleoperate an industrial robot arm via the Internet." The paper reveals that Mr. Goldberg and his graduate students "model frictional drag on the planchette with a constant magnitude and a direction opposite the current velocity of the planchette." So much for the mysterious spirit world.
Mr. Goldberg says his interest in such gadgetry is inherited. His father earned a metallurgy degree from Ohio State University and then opened a chrome-plating business in Bethlehem, Pa. But the work was dangerous, and his father set out to automate the plating process. "I would go down to the basement and work with him," remembers Mr. Goldberg, "as he reinvented the wheel a thousand times." Later, when he was an undergraduate at the University of Pennsylvania and his father was in a Philadelphia hospital dying of leukemia, they collaborated on a paper called "X-Y Interpolation Algorithms" that was published in Robotics Age.
The paper described an algorithm that lets a computer draw a circle without using large amounts of memory. A couple of years afterward, during a visit to a kibbutz in Israel while he was a graduate student in computer science at Carnegie Mellon University, Mr. Goldberg was able to put the algorithm to use in a machine that cut patterns in sheet metal. "It was an incredible moment," he says. "Something connected back to the real world in a way that could actually do some good."
By then, Mr. Goldberg was deeply involved in robotics. At Penn, he had studied with Ruzena Bajcsy, a computer-science professor in whose lab he had worked on sensors for robot fingers. (Ms. Bajcsy is now assistant director of the National Science Foundation's Directorate for Computer and Information Science and Engineering.) At Carnegie Mellon, he pursued a Ph.D. in robotics, and at the same time continued to discover interests outside computer science. He started spending time with students in Carnegie Mellon's College of Fine Arts, and he was introduced to a 1936 essay that still influences his thinking and his work -- Walter Benjamin's "The Work of Art in the Age of Mechanical Reproduction" (http://eserver.org/theory/work-of-art.html).
In 1991, he began teaching computer science at U.S.C. Almost immediately he became involved with the computer-painting exhibition -- a collaboration with Margaret Lazzari, a fine-arts professor -- that produced the Mulholland portrait in his office. The exhibition featured a computer-driven painting device -- which ran amuck during the opening, dragging its paint can across the canvas until Mr. Goldberg intervened.
"I was the first to admit that these were not great paintings," he says. But he adds that "the subject matter was trying to raise a commentary on what I was doing." The painting machine was "a symbol of the fragility of the infrastructure" -- and was just as vulnerable, he adds, as Mulholland's St. Francis Dam.
Mr. Goldberg says his years at U.S.C. were "hyperproductive," chiefly because he didn't have many friends in Los Angeles. "I would tell people I was a professor, and they would say, 'Wow -- you must be really smart. Why are you wasting your time?'" He tried his hand at a screenplay, but when he spent a semester at Berkeley as a visiting faculty member, he realized that what he really wanted to do was leave Southern California.
When he found out about an opening in the industrial-engineering department here -- "They were looking for someone with a computer-science flavor" -- it seemed like a perfect fit with his interests in robotics and in real-world applications for computing theory. He did not, however, mention his artworks in hiring interviews -- he had been warned at U.S.C. that it wouldn't help his career in engineering. "I felt like it was none of their business," he says now. "I felt like I wanted to be taken seriously as an engineer."
He's been a lot more public about his artwork since then, and is the founder of the popular Art, Technology, and Culture Colloquium here. "This campus has been very supportive of my artwork," he says. But his background in technology remains important, too. In his art, he says, "it's always been central that I'm critical of technology, particularly of the expectations in the popular press.
"But you don't want to stand up on a soap box," he notes. It's more effective "to develop a site that's somehow engaging but has a secondary message, like an element of warning" -- which was the point of the Mulholland exhibition. "I think," he adds, "that you can be the best critic from within."
http://chronicle.com
Section: Information Technology
Page: A45
