Someday, waiting in line for 30 minutes to attend a professor’s office hours or getting stuck with the graduate teaching fellow who loves to make midterms bleed with red ink may be a thing of the past, say computer developers who think machines could replace professors in the future.
Ben Goertzel, director of research at the Singularity Institute for Artificial Intelligence in Palo Alto, Calif., and Ray Kurzweil, a private inventor, teamed up with a group of researchers at NASA’s Ames Research Center to discuss the concept of
Singularity University, an institution where all the professors would be computers.
Goertzel and Kurzweil told The Chronicle of Higher Education that they believe computers will be smarter than humans by 2030, a moment dubbed “the singularity.” By that year,
Singularity University could be a real possibility, they said.
Kurzweil and Goertzel said virtual professors wouldn’t need tenure or salaries, and they could teach students individually
instead of in giant lecture halls. However, the duo added that computer professors couldn’t match individualized attention from humans.
University students voiced the same trepidations. “It would be a lot harder to ask questions or have things clarified,” said freshman Kayla Smith. She added that without help during office hours, students who don’t understand the material in the classroom would have to teach themselves outside of class.
Junior Nick Hiatt worried computer professors would lack human qualities, such as humor, compassion or personality. He said humor helps him stay engaged in a lecture, and respect for a professor helps motivate him to succeed.
“I hope they can find a way to incorporate the human aspect into it,” Hiatt said of Singularity University.
Both Hiatt and Smith said computers might raise logistical problems as well. They would eliminate jobs and opportunities, Hiatt said, by crowding out professors and graduate teaching fellows.
Hiatt raised another apt concern: “If there’s a power outage, you’re screwed.”
Smith said that with all the problems and drawbacks to computer professors, she didn’t think it would improve the learning or teaching process at all.
Whether students want computers to teach them or not, the phenomenon is already beginning. The University currently offers 29 classes under the title “Distance Learning,” where students take the class via Blackboard or other computer technology.
Professors use PowerPoint presentations to illustrate their lectures. Homework is often done online. WebAssign, a program that lets students do homework assignments via the Internet, is standard for University science classes. In many classes, quizzes are presented to the class on a projector. The students answer with a hand-held clicker, and the quiz is graded on the spot. Hiatt said his freshman physics class used that kind of computerized quiz.
The question at the heart of the computer versus human professor debate is whether computers can surpass humans in intelligence. Goertzel and Kurzweil think they can. Michal Young, a professor of computer and information science at the University, said people may be framing the question the wrong way.
“I’m not sure the question ‘is X smarter than Y?’ really makes any sense because X can be a lot smarter than Y in some ways and a lot less smart in others. Does playing chess well require one to be smart? If so, then computers are already smarter than people, but in other regards computers are not nearly as smart as cockroaches,” he wrote in an e-mail.
Young said that to determine whether computers can be smarter than humans, it’s imperative that humans know what they mean by “smart,” which could take longer than researchers are planning.
“The rapid improvement in computer power is irrelevant to this schedule,” Young wrote in the e-mail, “because the limiting factor is our puny understanding of intelligence.”
[email protected]