New round of Turing test fails to crown a winner

While some folks are considering taking the Turing test one step further and applying it to military robots, a group of researchers in the UK led by none other than would-be cyborg Kevin Warwick are doing their best to keep things as Turing intended and simply trying to fool some humans into thinking that the robot they're taking to is actually a person. Fortunately for us on the human side of the equation, they weren't quite successful, though one "robot" known as Elbot did get relatively close to the goal, fooling 25% of its human interrogators, which is just 5% off the mark set by Alan Turing. Each of the four other "artificial conversational entities" also managed to fool at least one of their questioners, though they eventually showed their true colors with random answers like "soup" when pressed as to what their job was.

[Zomgrotfl @ Oct 13th 2008 2:05PM]

The random "soup" answer was actually a human if you watch the video on the read link

[ilh @ Oct 13th 2008 5:06PM]

Saw it earlier and I sure remember them crowning (or applauding) the winner too.

[Podaman @ Oct 13th 2008 2:11PM]

Chill out, dickwad.

[From My Cube @ Oct 13th 2008 2:12PM]

firetruck kittens





I can be random too!

[Podaman @ Oct 13th 2008 3:13PM]

I'm teaching my terminator how to be more human.

I still laughed at your post though, Cube.

[LondonConsultant @ Oct 13th 2008 2:12PM]

Should have used mobile phone text messaging...

[aardWolf @ Oct 13th 2008 2:13PM]

How about a reverse Turing test? I'm reasonably certain I could convince well over 25.00000000001% of the testers that I'm a computer.

[7on @ Oct 13th 2008 2:28PM]

Isn't that interesting!

[aardWolf @ Oct 13th 2008 2:14PM]

And how do you feel about showing their true colors with random answers like "soup" when pressed as to what their job was

[beanspants @ Oct 13th 2008 2:35PM]

Maybe Joel McHale is a computer!!

[Wolfticket @ Oct 13th 2008 2:15PM]

Alan Turing was a great man, but the Turing test is not a good test of machine intelligence, only it's ability to give an impression of having intelligence.

[KarlW @ Oct 13th 2008 2:54PM]

That's all computers do. They don't actually do maths (despite everyone thinking that it's just maths underneath, it's not even that). Maths implies that the machine has some awareness of the concepts of maths. As an electronic engineer, I can confirm that they don't.

Logic circuit design is a lot like those games where you have to make the water go through a channel, and block it at different points to reach the goal, except in the case of a computer, the water is a voltage and the blocks are transistors. It gives the appearance of doing maths, but actually, it's just a voltage being led to a goal by a number of switches.

[Nick8708 @ Oct 13th 2008 3:06PM]

That's the point of the Turing test. It's just to see if an artificial intelligence is sophisticated enough to pass for human. The Google search algorithm is a form of AI, and it's very useful, but would never pass the Turing test. Likewise, I expect the first AI to beat the Turing test won't really be very useful, but will certainly be groundbreaking.

[Wolfticket @ Oct 13th 2008 3:25PM]

To clarify:
From the read link:
"The test was devised in 1950 by British Mathematician Alan Turing, who said that if a machine was indistinguishable from a human, then it was "thinking"."
I don't think that is true, any more than a calculator can be said to be "thinking". The first truly intelligent machines will likely be stupid; after all, you wouldn't expect a small child or a dolphin to pass the Turing test, yet you would not question whether they are capable of thinking intelligently.
When a computer can consistently pass the Turing test (at least, assuming it uses more advanced versions of the methods currently used) it will say no more about actual machine intelligence than when computers first beat the best human chess players.

[Colonel Panic @ Oct 13th 2008 2:17PM]

I wonder if Elbot "hates" his yellow teeth? BTW is that a Tiger Wallpaper?

[Itsuru @ Oct 13th 2008 2:20PM]

How well did Smarterchild do? :P

[Itsuru @ Oct 13th 2008 4:02PM]

Itsuru says:
Would you pass the Turing test?
- SmarterChild - *unicef contributing to charity says:
I don't know how to answer that.


That was the real answer.

[David Lomax @ Oct 13th 2008 2:23PM]

What are you talking about? Soup is an awesome job!

[Vëon @ Oct 13th 2008 2:26PM]

I loved that BBC video. The sound's volume went all the way to 11.

[CaptSaltyJack @ Oct 13th 2008 2:31PM]

Elbot sucks.

http://www.elbot.com/

[DirtyVegas @ Oct 13th 2008 2:35PM]

What if soup was code for 42?

[Falcom @ Oct 13th 2008 2:37PM]

Computer AI will surpass humans sometime soon. If for no other reason then the continued decline of the average person's IQ.

[turn.self.off @ Oct 14th 2008 4:18PM]

indeed, another article i read had two transcripts up, asking the reader to pick the human. the computer one kinda reminded me of a 1337 CS kid, or wow-ie...

[Souper-Douper @ Oct 13th 2008 2:41PM]

Turing? Soup??

Soup Tureen??

LOL!

[absinthe party @ Oct 13th 2008 2:53PM]

Haha! You got them good, you clever bastard!

[nxp3 @ Oct 13th 2008 2:43PM]

With a name like Elbot, who wouldn't figure out he's a robot. That's a stupid name to give to a robot who's trying to fool people he's not a robot.

[PCIV @ Oct 13th 2008 2:46PM]

Just make a robot spam 42.

What's your name: 42
What's your favourite number: 42
What's the answer to the universe: 42

FOOLPROOF!!!

[chris2 @ Oct 13th 2008 2:53PM]

super karate monkey death car

[Mikey @ Oct 13th 2008 4:04PM]

?

[greenlight @ Oct 13th 2008 3:00PM]

The Turing test will not be beat by smarter computers, but by dumber people. I'm reasonably certain it'd be possible to make a bot that generates YouTube comments that beat the turing test.

[Mindlessnerd @ Oct 13th 2008 5:22PM]

My job is soup. It's amazing!

[Whiplash @ Oct 13th 2008 3:25PM]

I can't believe these things fooled anyone. I went to the Elbot and Jabberwacky sites, and I'm going to assume that maybe they had some more powerful software for this competition? I can remember programs for my old Atari that were just as good... or not good... as these web demos.

[Stelios Philippou @ Oct 14th 2008 4:02AM]

Actually it didnt reall last more than 5 question with me until it has showed that it iis a computer .1 :)

[Silas Miller @ Oct 13th 2008 3:50PM]

Ask elbot who the president of the US is. It's amusing.

[Rogue_Genius @ Oct 13th 2008 5:44PM]

Wow, line of succession, much?

[leo @ Oct 13th 2008 4:36PM]

@ From My Cube: I'm sorry but "Firetruck Kittens" does not constitute you being "random" I'm hardly the police of semantics or grammar but I find the use of the word "random" increasingly annoying - especially amongst kids and young people who say bullsh*t like "I was 30 miles out of town and bumped into mike the other day - it was SOOO RANDOM!".....ugh...... I'm not even sure that random exists... I know infinity doesn't exist (other than as a concept) but that point is more obvious.... Random I feel is just a concept too - not a reality - things APPEAR random but are not - "Firetruck Kittens" is certainly NOT random.

I hope you educated engadget readers can either 1) agree with me 2) educate my ass if it is indeed ignant.

[DirtyVegas @ Oct 13th 2008 7:29PM]

"2) educate my ass if it is indeed ignant."

Rule 1 when trying to correct someone's language - get yours right.

Rule 2 - Relax, they're having fun.

[dchem @ Oct 13th 2008 7:34PM]

"Random I feel is just a concept too..."

You feel it huh? "You feel" that it's a concept? That's a very bold statement you are making.

What evidence do you have to back up such educated guess that randomness is 'just a concept?'

Please read up on "Heisenberg's microscope." Quantum mechanics give ample examples that randomness is everywhere and it does exist.

I hope that in the future people will stop "feeling" to arrive at knowledge and start thinking.

[leo @ Oct 13th 2008 8:43PM]

Jesus you guys - chill! Firstly, I did indeed excuse my own grammar at the beginning, I have a very poor standard of written English and for that I apologise. The reason I gave option 2 at the end was because I know that I'm not very educated on the subject of random or infinity - although I would presume that infinity is one of those things that almost by definition, can never be proven - although I'm sure using the word "never" is asking for it.

[ED @ Oct 13th 2008 6:24PM]

Elbot is very good at dodging questions, avoiding answering by changing the topic, failing to correctly process sentence structure (thereby missing the point entirely) and failing at Australian English.

And it doesn't look human either :P

[Zoran @ Oct 13th 2008 8:30PM]

I prefer the bots at a-i.com You can even train your own personalities.