Magazine article AI Magazine

Report on the First Conference on Artificial General Intelligence (AGI-08)

Magazine article AI Magazine

Report on the First Conference on Artificial General Intelligence (AGI-08)

Article excerpt

When the AI field began in the middle of the last century, its main focus was on the creation of computer programs displaying intelligence at the human level and ultimately beyond. This proved harder than anticipated, and over the following decades most research activity in the field shifted toward the creation of computer programs solving specific sorts of problems that humans find to require intelligence. This more task-focused "narrow AI" work that has come to dominate the field has yielded a variety of theoretical and practical successes, but there is no real consensus regarding the relation of this work with the original goal of creating powerful general intelligence at the human level and beyond. In the last few years, an increasing number of AI researchers have come to feel the time is ripe for a renewed focus on the original goals of the AI field; and this was the motivation for the convening of the First Conference on Artificial General Intelligence (AGI-08), which was held at the FedEx Institute at University of Memphis in early March 2008. Detailed information about the conference, including a complete proceedings, is available online for those who were not able to attend. (1)

The conference was characterized by an enthusiastic and energetic tone, and there was a feel of real excitement about having so many AGI-oriented researchers in one place at one time. There were more than 120 attendees (including roughly 40 presenters). The participants included not only AGI researchers from academe and industry, but also scientists and engineers from allied disciplines and a smattering of artists, business people, and other AGI enthusiasts, who providede a valued breadth to the discussions. The crowd was international, with nearly as many European as American participants.

The conference room was a futuristic setting called "The Zone," with a vague resemblance to the Star Trek bridge, including an excellent if mildly glitchy video system that, during Q&A sessions, displayed the questioner on a big screen in front of the room. This worked well with the unconventional format of the conference, which was divided into topical sessions, each one featuring 10-minute summary talks by researchers, followed by a lengthy moderated discussion session. Some of the sessions were markedly interdisciplinary; for instance, the session on AI in Virtual Worlds was chaired by Sibley Verbeck (CEO of Electric Sheep Company); and the session on neural nets was chaired by Randal Koene (a neuroscientist from Boston University).

A postconference workshop on "Sociocultural, Ethical, and Futurological Implications of AGI," (2) which drew about 60 participants, was introduced by noted futurist Natasha Vita-More and included some wide-ranging and passionate discussions.

Why Might the Time Be Ripe for AGI?

Given the checkered history of ambitious attempts at human-level AGI, it may justly be wondered whether the time is yet ripe for another volley of such attempts. This was one of the main topics of discussion in the opening session of the conference. Following a general introduction by conference chair Stan Franklin (University of Memphis), Ben Goertzel gave a talk on "The Past, Present, and Future of Artificial General Intelligence." While the researchers in the audience displayed a healthy level of disagreement regarding this topic, the overall opinion was that it does make sense, at present, to focus considerable resources and energy directly on the AGI problem, as distinct from more task-specific, narrowly-focused AI initiatives.

As an example of the issues raised in the discussion following Goertzel's introductory talk, Hugo de Garis asked Goertzel: "People have been trying for 60 years to produce human-level AI, and have failed. Why do you think the time is ripe to try again? What has changed, if anything?"

Goertzel responded: "I think four things have changed: (1) Computers are far faster now, and far more easily networkable. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.