Monday, June 09, 2014

« Second Life Designers are Harnessing the Power of the Animated Gif for Eye-Catching Virtual Product Ads | Main | This Nude Descending a Staircase is Not the Nude You're Looking For (But You Should Look Anyway) »

Chatbot Avatar "Eugene Goostman" Passes Turing Test

As you might have read over the weekend, the 65 year old Turing Test was just passed for the very first time -- but not as often reported, via a "supercomputer", but a chatbot avatar personality named "Eugene Goostman" -- more on him in this video from last year. I think it's fair to call Eugene an avatar since he's designed with a very specific personality (a smartass teen boy). You should typically be able to chat with Eugene at this link, but the site seems to be down due to traffic. I wonder if his chatbot technology can be introduced into virtual worlds. Blue Mars featured a chatbot which scored high on the Turing Test. Then there's chatbots like "Social Autopoiesis" in SL , which I once had a charming, foul-mouthed conversation with:

We totally passed the Turing Test for F Bombs. Update, 6:03pm: On the other hand, maybe Eugene never really passed the Turing Test and maybe it doesn't quite matter if he even did.

Please sharehare this post:


Feed You can follow this conversation by subscribing to the comment feed for this post.


The results are suspect.

The bot was based on a real boy who could not speak english. It was also not based on Turing's test guidelines but Reading University's interpretation of his guidelines. Come on! Do none of these people have intellectual honesty anymore? They are going to need plenty of it if they even hope to create an AI (which I think will NEVER occur).

Arcadia Codesmith

Turing never intended the test to be a practical measure of AI, and it's not regarded by AI researchers as much more than a novelty.

AI is here, and it's only getting more and more sophisticated. Denying that AI will ever be developed at this point in history is like rejecting the laws of gravity.

Strong AI, which achieves or exceeds human standards of sapience, is not here yet, but I would be shocked if it wasn't available before the end of this century.

For games and virtual worlds, a chatbot that is Turing compliant (or close to it) could be a very significant advance in the development of smarter NPCs and just possibly could pull us out of the "quest on rails" rut that the MMO industry is currently stuck in.



The Test is regarded as a novelty because they can't pass it without fudging. And the more they fudge the more they change the rules on the test.

Having a computer throw out pre-recorded speech that is halfway relevant is not the same as a real human to human conversation. Not to mention that they had to dumb down the AI so that it isn't even an adult to adult question answer.

Fact is...a lot of people are wasting money on this delusion.

Nathan Hopkins

Are you going to correct your story?

Extropia DaSilva

Why would artificial intelligence never occur? It is not as if intelligence is impossible, for if it were nature could never have evolved animals like humans and dolphins and dogs which have it, would it?

One only has to imagine how useful it would be to have machines which can usefully plan, act, and adapt to see the commercial pressures for developing intelligence in machines, so not only is it quite possible, there is also a massive incentive.

I suspect you are confusing intelligence with free will or consciousness. Maybe tech will not have such attributes but robots that show true intelligence and freedom to learn and adapt within certain constraints? It is a near certainty.

Extropia DaSilva

Hamlet asked if it would matter if and when a chatbot really displays the ability to converse in a manner that demonstrates true fluency in that language and understanding of the conversation.

The answer is that skeptics of machine intelligence will insist it does not, for we know from experience how they respond to inexorable progress of technology:

"Machines will NEVER do X"

(then a machine is invented which does do X)

"But machines will never do X as well as the top human experts"

(The machines completely trounce the top expert)

"Oh, well, X was never of any real importance, anyway".

Post a comment

If you have a TypeKey or TypePad account, please Sign In.