Public meeting will re-examine future of artificial intelligence

Written by Tom Abate - 7 September 2007

For decades, scientists and writers have imagined a future with walking, talking robots that could do everything from cooking your eggs to enslaving your planet. Trouble is, this fabled artificial intelligence has never happened. But this weekend, more than 700 scientists and tech industry leaders will gather at San Francisco's Palace of Fine Arts Theatre to plan for the day - still decades away - when computers start improving themselves without the approval of their former masters. Participants wonder whether this will yield the kindly Commander Data of "Star Trek: The Next Generation" fame or the mob of killer machines that attempted a world takeover in the movie "I, Robot."

"The history of technology tells us that inventions can be used or misused for good or evil. It could be that an Orwellian state could use this technology, or it could lead to a world with more accountability and transparency," said technology financier Peter Thiel, a principal backer of the two-day event called "The Singularity Summit: AI and the Future of Humanity."

The Singularity is the term used to describe this anticipated - or feared - day when machines become smart and perhaps ambitious enough to reprogram themselves. This weekend's gathering expands on a similar event held last year at Stanford University.
Thiel, who holds philosophy and law degrees from Stanford, co-founded PayPal and sold it to eBay in 2002 for $1.5 billion. In addition to running Clarium Capital Management, the hedge fund he founded in San Francisco, Thiel, 39, is using his wealth and celebrity to raise public awareness of the stakes surrounding artificial intelligence, or AI.

So, why should society take AI seriously now when its promoters have oversold it so far?
"The pendulum has swung too far, and people now underestimate it," said Thiel, arguing that recent advances in computer hardware, software, cognitive science and computer networking have created a technological primordial ooze. Moreover, these primitive AI systems have already made themselves useful parts of everyday life.

"Google, to a certain extent, you could characterize as a limited artificial intelligence," said Tyler Emerson, executive director of the Singularity Institute for Artificial Intelligence, the Palo Alto nonprofit group sponsoring the weekend summit with financial support from Thiel.

As Emerson explained, computer scientists now freely admit that they vastly underestimated the complexity of human intelligence when they first defined AI as an all-knowing, smarter-than-human system that could do everything from calculating the trajectories of planet-killing asteroids to composing an opera.

But while that general-purpose vision of AI has proved elusive, Emerson said technologists have gradually built and deployed special-purpose artificial assistants that we already take for granted.

In addition to search engines, he cited the way nonhuman characters in online video games react to their human counterparts. He noted that 10 years ago, Deep Blue, the IBM computer, whipped world chess champion Garry Kasparov. And he acknowledged advances in robotics exemplified by Stanley, the Volkswagen that drove itself across the Mojave Desert in 2005 using onboard sensors and software developed by computer scientists at Stanford.

Recognizing the confluence of these and other developments, a growing number of computer scientists, ethicists, industrialists and forward-thinkers believe that, far from being improbable, machine intelligence seems to be evolving, if that biological term is adaptable to human inventions. And given that a strictly laissez-faire approach could cause a potentially evil genie to escape from the bottle, perhaps some forethought could coax technology in a more beneficial direction.

"There is still time to affect this," said Thiel. "There are choices to be made." Toward this end, the Singularity Summit will bring together experts in computer science, nanotechnology and related fields, including:
-- MIT robotics Professor Rodney Brooks, co-founder and chief technology officer of iRobot, the Massachusetts firm that sells self-directed machines ranging from automated vacuum cleaners to mechanical scouts and pack mules for the military.
-- Peter Norvig, director of research at Google and author of "Artificial Intelligence: A Modern Approach," the foremost textbook in the field.
-- Nanotechnology investor Steve Jurvetson and Foresight Nanotech Institute co-founder Christine Peterson.
According to Emerson, it was Peterson, the nanotech maven, who put Thiel in touch with the Singularity Institute. Peterson, who holds regular salons for Silicon Valley thinkers, invited Thiel to dine with South Bay mathematician Eliezer Yudkowsky, who co-founded the Singularity Institute in 2000. Thiel and Yudkowsky hit it off, and soon the financier was helping support this AI ethics outfit.

"It will be a watershed event," Emerson said, imaging what it would be like as machines start acting independently, especially to add powers their human creators had not envisioned. "We're dealing with the most powerful process we know at this time, which is the power of human intelligence," he said. "Intelligence is not a tool. It is something that creates tools."

Thiel declined to disclose how much money he has put into the institute, but Emerson said it was sufficient to boost the organization's profile such as with the coming conference. Last year's event at Stanford was free and drew 1,300 people. It was headlined by inventor Ray Kurzweil, author of "The Singularity is Near," the unofficial bible of this movement. The Palace of Fine Arts Theatre seats 962, Emerson said, but given that this time the group is charging $50 for the two-day event, organizers would be thrilled to sell out.

But however big or little the splash they make, Thiel and his band of scientific soul-searchers want to call public attention to where technological currents seem to be heading, and to presage both the promise and peril that may lie ahead.

"There's a lot of debate about whether computers can think," Thiel said. "It's good for humans to use their brains to think about the future every once in a while, and that's what the Singularity Summit is about."

Available: [Online] http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2007/09/07/MNK8RUU7J.DTL&feed=rss.business