April 29, 2003
Taking the test

Salon carried a really great story 2 months ago on Hugh Loebner and the Loebner prize. A madman it seems from the article. The concept of a Turing Test of an artificially intelligent machine (fooling an unknowing guinea pig into belieiving that the AI is real through conversation) will be known to many people. The Loebner prize is the embarrasing not-too-serious competion to actually run a Turing Test. The best machines entering get a 'best this year' award, but to actually claim the prize they have to actually fool a living person. The best machines are about as far from that as Leonardo da Vinci was from building an actual aeroplane.

Loebner, it turns out, is a rather strange person and fun to read about. The writer, John Sundman, is a little too happy to root for the little guy, against the establishment. In doing so he picks at several of my personal heroes (Marvin Minsky and Daniel Dennett) on what I consider completely unfair grounds. Both are accused of being establishment scientists locked up in dogma and absurd theory, and in these particular cases nothing could be further from the truth. Minsky and Dennett are to be commended for their extremely hands-on approach to doing science and philosophy respectively. Furthermore they have written very accessible books about their respective fields.

Minsky's society of mind is an extremely practical guess at some of the architectures that will be necessary to build actual intelligence of any kind. It is however not theory-free as Sundman and the infamous Richard Wallace (af A.L.I.C.E. fame - and quoted in the article) would like.
A.L.I.C.E will never win the Loebner prize - it just doesn't learn enough from the conversation even though one can seed it with very convincing replies.

Dennets consciousness book I haven't read. But 'Darwin's Dangerous Idea', while challenging reading is also one of the best books I have read in a while.

Sundman manages to give the impression that these are medievel progress hating nay sayers whereas they are in fact hardcore believers in manmade complex machines, call them intelligent or not.

Posted by Claus at April 29, 2003 12:08 AM
Comments (post your own)
Help the campaign to stomp out Warnock's Dilemma. Post a comment.

Email Address:


Type the characters you see in the picture above.

(note to spammers: Comments are audited as well. Your spam will never make it onto my weblog, no need to automate against this form)


Remember info?