AI @ 50: Can It Be?

Can it be? Is AI is really 50-years old? Well, almost. It will be exactly 50-years old at the end of August, but the festivities are underway sooner than that.

Okay, we acknowledged the anniversary earlier this year in the January 2006 issue of Dr. Dobb's. It was in 1956 that John McCarthy, Marvin Minsky (who along with McCarthy founded MIT's AI lab), IBM's Nathaniel Rochester, and Bell Lab's C.E. Shannon presented "A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence" at the Dartmouth Artificial Intelligence Conference. The conference and the project, convened with the goal of creating truly intelligent machines, established AI as a unique field of study within computer science.

To celebrate the occasion, Dartmouth will be throwing a party. Okay, a conference. Officially entitled "Dartmouth Artificial Intelligence Conference: The Next Fifty Years," the AI@50 conference will convene on the Dartmouth campus July 13-15 and include talks by McCarthy, Minsky, Rod Brooks, Stuart Russell, Nils Nilsson, and other luminaries in the field of AI.

According to conference director James Moor, there will likely be considerable debate about the future direction that AI should take. He says that the plan for the conference is not only to honor the past and assess present accomplishments, but also to help seed ideas for future artificial intelligence research.

"AI has come a long way in 50 years, and it has a bright future. Although computers may never replace us, smart machines will be prevalent in our environment, and may someday be even implanted in us. The future of artificial intelligence deserves careful and sustained scrutiny."

He considers the field on two levels:

• One is the applied computing level that involves the development of "expert systems," for such areas as spectrographic analysis, stock market patterns, and medical diagnostics; and the development of robotic systems for driving cars as well as software for searching the Internet.
• The other is the philosophical level that tries to answer a question such as, what is the nature of intelligence?

He thinks that bridging these two levels is challenging, but very useful in developing a fuller understanding of minds. Science forces philosophy to be more empirical and philosophy forces science to be more reflective.

"The initial hope was that AI could do much more than has actually panned out in 50 years," he says. "Language use and translation by machine, for example, was once expected to be quite easy. Turns out that language use requires extensive knowledge of how the world works. This complex knowledge usually assumed in ordinary conversation, which is easy for us, but it's difficult to teach a computer voice inflection, social and cultural idiosyncrasies, as well as a multitude of social contexts."

Posted by Jon Erickson at 09:38 AM  Permalink