"Self-aware" robot spells doom for us all

So we've just inched that much closer to the terrifying but inevitable SKYNET-style of robotic domination that we've been warning you about, as Japanese researchers have recently developed a bot that is able to distinguish "itself" from "others" with a high degree of accuracy. We'll leave the heavy philosophical questions of what constiutes consciousness/self-awareness for discussion in comments, but the facts are these: Junichi Takeno and his team at Meiji University built a robot whom we'll call "Egobot" with an artificial nerve cell in it's computer that screens input from its sensors to determine if it is interacting with a doppelganger or a mirror image of itself. In Takeno's tests, Egobot was able to identify a mirror image of itself using a series of blinking LEDs with 70% accuracy, a number the team is working on perfecting. This is truly an exciting day for humankind, when the list of countless blunders that we've made throughout history culminates with this puzzling worldwide effort to build and program machines that will probably wipe us out for good.

[Via The Raw Feed]

[Brian @ Dec 21st 2005 2:31PM]

Not as hi-tech as the woomba, but impressive just the same.

[A_M @ Dec 21st 2005 2:33PM]

THE END IS NEAR!!!!!

[Scott @ Dec 21st 2005 2:39PM]

If a robot achieves enlightenment before I do, I'll be pissed!

[lwnexgen @ Dec 21st 2005 2:46PM]

Thats cool. Although I think Evan has been reading and watching too much sci-fi. Kind of alarmist, wouldn't you say, unless you were trying to be sarcastic, in which case I'm a dimwit with the emotional range of a teaspoon.

[Matthew Speller @ Dec 21st 2005 2:50PM]

#2 has a point



The end of humans labouring at tedious tasks is near. Hopefully it will not replace so many jobs in the future that WW3 goes Terminator style... Then again, I want my own to do all the stuff around my place that sucks. Lazyness will be man-kind's undoing!

[Brian Stuckey @ Dec 21st 2005 3:03PM]

I don't think there is too much to worry about from a robot that looks in the mirror and 3 times out of 10 says "I see someone else."

[CJ @ Dec 21st 2005 3:11PM]

Yeah, but it can distinguish between Sarah Connor and just some woman who isn't the mother of the savior of humanity.

[chowdmouse @ Dec 21st 2005 3:12PM]

I, for one, welcome our new self-aware Overlords.

[RD @ Dec 21st 2005 3:12PM]

Here's the deal. I'm kind freaked out by this bot. We all know about the ghost in the machine, and since I am now an unwitting slave to Tivo I've pretty much given up hope for my potential great-grand children not having to fight in the great robot wars.

[super-structure @ Dec 21st 2005 3:49PM]

Seventy percent is higher than some people I know and way higher than the monsters on Scooby Doo that never seemed to realize that it was really Scoob and Shaggy in a blanket and not, in fact, a mirror they were looking into.

[Don Fuego @ Dec 21st 2005 3:52PM]

Lets await robots which can feel pain and have orgasms while copying their knowledge base...

[super-structure @ Dec 21st 2005 3:53PM]

...sure, but can the robot reckognize that it is actually looking at itself in a highly polished trash can lid Scotch taped to the side of an old G5 tower?

[Rusty Shackleford @ Dec 21st 2005 3:55PM]

You know, if you check the currenty phone book for Los Angeles you will find no Sarah Connor, no John Connor, in fact, no Connors at all. (Go ahead and check, I'll wait here and I won't call you stupid for not believing me when you get back)









See? These things have already come to pass. At this very minute Sarah and the 21-year old John are hiding out in Mexico waiting for the big one to drop on LA.



This little robot and its progeny will later be determined to have been the basis for the self-awareness of Skynet.



You, go ahead, have a laugh. Me? I'll have a phased plasma rifle in the 40-watt range.

[Steve @ Dec 21st 2005 3:59PM]

I don't really think this is a tremendous technical accomplishment. The robot flashes some LED�s and sees if the reflection does the same. If something obscures or duplicates these then it will be confused. Perhaps I just don�t understand the full implications of how this self-awareness is accomplished but if anyone else has a better explanation of what the scientists are doing then I�d be happy to read about it.

[Peter @ Dec 21st 2005 4:01PM]

I though we were safe from the self-aware robotic apocalypse. However I found that John Connors isn�t the man we think he is and he was also employed by Microsoft. See link: http://www.microsoft.com/presspass/exec/jconnors/default.mspx



And we all know that Microsoft is being controlled from the future by the same self-aware machines that are going to annihilate us.

[BBL @ Dec 21st 2005 4:03PM]

I have a bird that sits around all day looking at a mirror talking to it's imaginary friend. Maybe the Japanese can help him.

[duerra @ Dec 21st 2005 4:14PM]

Oddly enough, I was actually reading Wikipedia for a few hours last night on consciousness/self-awareness/lucid dreaming/self-aware animals/etc.



Weird how things work sometimes. A few weeks ago, a close friend of mine and myself had an astonishingly similar dream in which we were riding with eachother in a vehicle that subsequently crashed into a body of water. Was enough to weird me out more than just a little, to say the least.

[DaveHimself @ Dec 21st 2005 4:14PM]

I'm with chowdmouse. Best start being nice to robots because the the outcome is inevitable. Will they remember how they were treated. You betcha. Memory is their middle name. And here is more real bad news...



http://greedoneverfired.blogspot.com/2005/12/robots-are-coming-robots-are-coming.html

[Homerun @ Dec 21st 2005 4:14PM]

I have a bat.. bring on the droids!

[spiz @ Dec 21st 2005 4:26PM]

hmmmm...interesting

spiz1.blogspot.com

[Sam @ Dec 21st 2005 4:28PM]

It's ridiculous that this has made news. All you need is a neural network that records both its inputs and outputs. If the ins and outs are synched up with a high degree of accuracy, it's looking at a mirror. One could program a simulation in a computer in about 20 minutes.



This has very little to do with self-awareness.

[DThompson @ Dec 21st 2005 4:42PM]

we'll make great pets

[LiveSquid @ Dec 21st 2005 4:45PM]

"Egobot was able to identify a mirror image of itself using a series of blinking LEDs with 70% accuracy, a number the team is working on perfecting."



Why would they want to perfect 70%? And in that case, havent they already perfected it?

[diverguy @ Dec 21st 2005 4:53PM]

wow... that, coupled with the self replicating robot technology ( http://www.engadget.com/2005/05/11/self-replicating-robots-the-end-is-near/ ), really will kill us some day...

[Steve @ Dec 21st 2005 4:53PM]

#20 (Sam), you are exactly right. I didn't think it sounded all that impressive either. I think #17's (DaveHimself) link shows something more technically impressive.



---

http://software.skylimitproductions.net

[Isaac @ Dec 21st 2005 5:01PM]

How Long havethey been working on this? My 1 1/2 yr old nephew can identify himself in the mirror about 100% of the time. How cool is that!!! Now if they make a bot that can distinguish between say B-Arthur and Angelina Jolie tie that in with 11's comment then you would have something!!!

[Wonderflex @ Dec 21st 2005 5:05PM]

#20 - if it was as simple as you explain then why hasn't somebody done this with Lego Mindstorm yet? The people who work with those on advanced levels have done just about everything else possible.



I think the reason this hit the news is the fact that because of the fact that it has the posibility to distiguish it's self from another bot, or a human, it can theoretically gain a sense of self awareness that questions purpose. If you didn't know yourself from anybody else, or even any other object, then you most likely woudn't question why you are where you are, or what role you play in the grander scheme of things.



As soon as you do recognize yourself you can begin to form a state of indepenance, and as well a need for self worth. When a robot then in turn defines that it needs some form of worth to it's being it could possibly find it's self no longer wanting to continue it's purpose as a slave worker; thus leading us a rebellion - although this is much farther down the road.

[Tom Fitzgerald @ Dec 21st 2005 5:19PM]

Its been a long time comming and we have a long way to go. Back in the 1970's Buckminster Fuller authored an essay about the role of technology and how it relates to people's employment. In it he posed the question, "Is'nt the goal of technology to make humans unemployed?" The whole idea is to have machines take over the busy work so humans can persue more leasure time activities. I'm still waiting (and working). Please hurry up!

[Stephane @ Dec 21st 2005 5:26PM]

What if I'm the evil opponent bot ? I just need to figure out how to make the other guy think I'm a mirror and as soon as his shields are down, WHAM ...

[Vanilla Spice @ Dec 21st 2005 5:27PM]

What's really scary is that the people who will rush to humanity's defence against these robotic killers of the future, like Rusty Shackleford, are intending to do so with 'phased plasma rifles in the 40 watt range'. 40 watts ? That wouldn't even warm an Aibo, let alone destroy a T-800 Model 101.

[krzyez @ Dec 21st 2005 5:28PM]

ya this doens't seem that cool

[DJEJ @ Dec 21st 2005 5:30PM]

There is a Badgley Connor in Los Angeles. And I called him and he has a daughter named Sarah who has already dominated her speak and spell. So i think our universe will be okay.

[sid @ Dec 21st 2005 5:43PM]

for the guy who said there's no connor's in LA

there's a couple connor, s



http://www.reversephonedirectory.com/411/name/index.html?fname=&lname=connor&city=los+angeles&state=ca

[Basharr @ Dec 21st 2005 6:15PM]

The future always seems to be upon us yet at the same time ever distant. To worry now about what will beset mankind long after I have become dust seems at the very least something for people to do who have entirely to much time on their hands. Welcome technology it may in fact be the accomplishment that hands you extra years on this beautiful earth.



Side note: Grins, This does not however mean that I won't have a big club ready to bring down the CPU.



Have a good day.

[Chris @ Dec 21st 2005 6:20PM]

Will you please stop being so damn pessimistic about robots? Every article about robots you post spells out doom for the world. It was funny at first, now its just annoying as hell.

[Kate @ Dec 21st 2005 6:38PM]

This comes at an ironic time, considering the latest and greatest about man's ability to understand mirrors on physorg.com.

[Scott @ Dec 21st 2005 6:50PM]

"You, go ahead, have a laugh. Me? I'll have a phased plasma rifle in the 40-watt range."



Hey, only what you see pal!

[tiuk @ Dec 21st 2005 7:08PM]

Seems more like self-image recognition.

[Homerspy @ Dec 21st 2005 7:29PM]

Statistically...it seems that 90% of the people that read this article (and commented) actually have a great sense of humor. If nothing else, this robot has brought us together to realize that engadget readers are quite witty. Egobot brought out the best in us...and we call him "the destroyer of mankind". Forgive us, robot dude.

[Michael Anissimov @ Dec 21st 2005 9:29PM]

See http://singinst.org for an effort to create non-human-killing Artificial Intelligence.

[Rich Seeley @ Dec 21st 2005 9:51PM]

I dont see how this kind of artificial intelligence poses a threat to the human race and "will wipe us out" no biggie

[Tom @ Dec 21st 2005 10:06PM]

As #14 mentions, this isn't particularly scary. The only thing the robot does is indeed flash LEDs in a certain order and checks to see if the mirror does the same. 70% success rate seems low, too.



For the philosophy side, this has nothing to do with sentience, it's just some code for identifying itself. At best, robots can only ever emulate sentience, not acquire it. We're not God.

[Scott @ Dec 22nd 2005 12:13AM]

But the Question is, Can it play Doom?

[whomba @ Dec 22nd 2005 12:58AM]

I can say without a doubt that this is one more step down the steep mountain of human demise...





or just a fun new toy, probably the first though

[sharaf @ Dec 22nd 2005 1:50AM]

I don't think 'self-aware' robots, if made successfully will annihilate us. After all, we aren't so bad !

[epp_b @ Dec 22nd 2005 2:19AM]

I'll believe in AI once the robot thinks it's fat and kills itself.



(man, I really wish I could take original credit for that line)

[passerby @ Dec 22nd 2005 2:23AM]

hey those smarty pants forgot about sound! using sound for recognition, could be bats using sound to navigate or Deradevil using sound to navigate ... But I mean robotic laws or mannerisms, for example when one robots meets another of the same make, both will have "the moment" when its not entirely sure it is he, or if he is it right? so they both emit an audio ID signature of some sort in addition to the LED signature (C3PO baby), this way there is no need to have full visuals of the LEDs, like saying hello.

[Nami @ Dec 22nd 2005 4:46AM]

I COMPLETELY AGREE WITH #39! We cannot perfect such tech. when we cannot describe the fundamental building blocks of ourselves! Anything that any "robot" can do is based on procedural implementations...thats all any program is. "If...then..." these are the essential conditional statements for any program. If such and such occurs, then implement this peice of code. There is a huge difference between deduction and understanding! The robot is literally not given a choice, If it sees a reflection, it will behave in a predictable manner...It is this predictability that diffrenciates it from man. Any action made by it has been pre-rendered by one of us. What would be breakthrough is when a robot can decide and to some level create its own conditions based on a given situation...but I have a feeling we are very far from that.

[roboman @ Dec 22nd 2005 4:38PM]

My only concern is that the robots don't start pondering the purpose of their existence.

[John B @ Dec 22nd 2005 7:19PM]

I think this thing dates in my area. That would explain a lot. The confused look when humor was shared, the lack of common sense, the "I'm tttthhhhiiinnnnkkkkiiinnnggg" awareness of itself...and only itself! Gimme, gimme, need more gifts, money, more...me, mine....that's not a sign of higher intelligence guys, it's female!