MIT Scientists Create Robotic Sea Life

Junior Barns writes "This article on the BBC News site reports on the development of a robot that imitates primitive life forms. This project led by researchers from the robotic life group at the MIT media lab is intended to study how people will try to interact with and relate to an "alien" creature that seems organic but is not anthropomorphic. Let's just hope no one tries to kill and eat it."

Comments

yum...

[skydude_20]

Let's just hope no one tries to kill and eat it.

but how else will we be able to know if an "alien" creature tastes like chicken?

The real link

[unsinged int]

Here is a working link to the story. And a working BBC link.

To robots, tires are like candy!

[tcd004]

Chains, nails, you can chew them up with your titanium teeth and spit them out like bullets!

Sorry, this just spawned memories of my favorite SeaLab 2021

Visit
tcd004

Prevalence of A.I. stories = AI /. topic needed?

[wackybrit]

In the past few weeks we've had a lot of stories relating to Artificial Intelligence. Yesterday we had that game characters one, we had the computer that built an oscillator, news about 'Ai' and 'Cyc', and so forth.

AI is definitely becoming a reality. Everyone was interested in AI ten or twenty years ago, but it's NOW that things are starting to really happen. The technology is here.

So why doesn't Slashdot have an 'AI' topic? I think it's time we had one here, as AI is clearly becoming a popular topic on the site.

P.S. I believe in this enough that I'm willing to burn some karma by posting this almost offtopic post, which will probably be modded down as such. Mod me up if you agree, or mod me down if you think it's a bad idea.

Wow...Nice Job

[unsinged int]

That first robot is pretty ugly, but the one in the second picture looks great to me.

You may stop the research, results are here.

[jukal]

" intended to study how people will try to interact with and relate to an 'alien' creature that seems organic but is not anthropomorphic"

Based on a empirical research conducted amongst 42^42 volunteering /. readers an average humanoid reacts by first looking up some words from the thesaurus continued with anthropomorphous expressions like *pause* *sigh* and *Huh!*

Re:You may stop the research, results are here.

[i0lanthe]

My first thought was "Someone's been playing too much Seaman"...

Access problems

[K.]

If you get 403s, try clearing your BBC cookies and going via the front page, answering yes to the are you from the uk question. Worked for me.

I doubt it was slashdot wot done it too, more likely someone fucked up file or CMS permissions and hasn't noticed coz of said cookie being set to "yes" on all BBC boxes.

Re:Access problems

[fiiz]

That's right.

The "non-UK" version of the BBC has been down for a few hours, it's definitely not ben /.ed, there's a silly permissions problem.

And, as you say, the "UK" version works fine.

Re:Prevalence of A.I. stories = AI /. topic needed

[Elwood P Dowd]

Better yet, reply if you think it's a bad idea.

For example, I think it's a bad idea.

Mostly because AI is generally an argument topic. It'd be like having an "Evolution" topic. Whenever anyone brings up AI, you get twenty highly rated posts talking about "Why don't the researchers see that AI would be easy if only they [insert poorly thought-out idea here]". And then you get like three voices of reason explaining why none of this is groundbreaking.

Er... I just made your point. If we had an A.I. topic, then everyone with my complaint could filter out the flames....

I guess I'm just trying to say there shouldn't necessarily be *more* AI stories.

Re:Prevalence of A.I. stories = AI /. topic needed

[Kizzle]

Don't worry, as long as you say your off topic there is no way you can be modded down.

Tomorrow in the news:

[cscx]

MIT Scientists Create Robotic Girlfriend... story at 11.

USS Virginia, somewhere under the Northern Pacific

[Subcarrier]

First officer (staring intently through the view finder): "Captain, I think we may have a slight problem."

Captain: "What is it?"

First officer (in a strangely strangled voice): "I think you had better see for yourself."

Captain (taking over the view finder): "What the hell is THAT? Is that thing trying to...?"

First officer (face now twitching almost uncontrollably): "I believe so, sir."

Captain: "But... It's humping my..."

First officer (gasping for breath): "Some of us like our bitches big, sir."

Captain (dazed): "My mission..."

First officer (now laughing outright): "We could always abort, sir."

AI icon?

[karm13]

let me suggest HALs eye as an icon.
how would you express AI as an icon? maybe that's why it isn't here yet.

Re:AI icon?

[Saeger]

I would suggest that AI go under an "Emergence" topic instead (since AI must be evolved bottom-up, not designed top-down :), along with ALife, GA, GP, NN's, etc.

Maybe use an anthill for an icon to depict complexity naturally emerging from simplicity? Or maybe a few cells from conways game of life?

--

Sounds familiar

[Lars T.]

[...]how people will try to interact with and relate to an "alien" creature that seems organic but is not anthropomorphic.

So they build robotic Sea Monkeys? (Note the "not anthropomorphic" bit and how the SMs were supposed to look like)

Slightly Offtopic

[Cyno01]

Has anyone seen that new show on the Discovery channel, Depth Charge. Its kinda like robot wars, except its under water, i've seen one episode, it was pretty cool, its a little bit slower paced than battle bots, but there seem to be less restrictions on weapons, bots last night had plasma cutting torches and bang sticks and other cool things.

Re:Prevalence of A.I. stories = AI /. topic needed

[wackybrit]

Mostly I agree with you. The discussions on AI on Slashdot really haven't been of the highest quality.. but then again, neither have the stories.

I do believe, however, that AI is experiencing exponential growth at the moment, and that AI is becoming what people twenty years dreamt of. Sure, we have no HAL yet (another ten years, I bet ya) but there are going to be enough stories coming along that are directly related to AI that I think a designated topic would be useful.

Then again, I also believe topics like 'PHP', 'Perl' and 'C++' should also be culled, and instead use 'Programming'.. but hey.

And anyway.. how comes there's no area on Slashdot where we can actually discuss the workings of Slashdot? A 'MetaSlash', if you will? Journals are good, but there are none that are particularly popular for this type of discussion.

Suggestions?

Return of the Creeping Eliza

[WillWare]

The Media Lab website has the following blurb:

we are building a robotic sea anemone-like creature... The fundamental questions explored by this work are:

• What interaction skills must a robot have for people to feel they are in communication with it?
• Do these skills allow people to anthropomorphize it and in what way?
• How do these skills influence a person's opinion of how intelligent the robot is, how alive it is, and what other abilities it might have?

This is Wiezenbaum's Eliza experiment. The researchers are hoping people will overestimate its intelligence. The measure of success here will be how well they mislead people. How much new insight can we expect from this version?

These guys will build a comlpex widget with sensors and motors that works underwater for extended periods. Why not make it do something clever, useful, unexpected? I know, I know, that's AI, and we don't do no stinkin AI no more. But it would impress me a lot more.

Saw it at SIGGRAPH

[mistermund]

The Public Anemone was presented at SIGGRAPH '02 in San Antonio, about a month ago - that's where the pictures in the article were taken. The exhibit was in the Emerging Technologies area. I visited the exhibit almost daily (reactive robotics is an area of interest), and spent some time observing both the exhibit and people's reactions.

The Media Lab students explained that it was an experiment in social interaction - but how people react with something that doesn't have a face, or a voice. In a way, it's easier to create a creature that doesn't have to synthesize speech, etc. At the same time, it's much more difficult to elicit a reaction from people when they can't interact the same way that they do with other humans.

The Public Anemone had two main forms of reaction that I could make out - shrinking back from people who reached out toward it, and tracking faces. (With the assistance of dual stereo cameras in the back wall.) The exhibit was more like a terrarium than an aquarium (as the BBC article mentions), but the creature had a silicone skin which allowed it to play in the small pond and waterfall without shorting. During the day cycle, the anemone interacts with guests. During the night mode, the anemone goes to sleep and guests can interact with other fiber-optic anemones (that also shrink away) and drum on gemstones embedded in the surface of the exhibit. The exhibit certainly looked cool, with fiber optics, a soundtrack, and changing colored stones (using ColorKinetics lights), but the interaction left something to be desired. Almost all the people I observed in the exhibit did the typical museum "Oh, that's nice, let's look at it for a few minutes." Almost no-one tried to interact unless prompted to by the media lab representative that was standing there, describing what was going on. Nobody that I saw tried to play with the face tracking abilities of the robot.

Cynthia Brazeal(the person in the second pic) is more commonly known for her work on Cog & Kismet. (Pic)

IMHO, The coolest project in this area is Doc Beardsley, by the Entertainment Technology program at Carnegie Mellon. Here's an article at Discover Magazine. Interaction with Doc emphasizes fun over artificial intelligence.

I have more pics of the Anemone from Siggraph. If anyone wants to post them somewhere where they can stand the slashdotting, send email to mistermund@yahoo.com

Rod Brooks and the Toy Factory

[Animats]

This sort of thing has been seen before. Rod Brooks, head of the MIT AI Lab, has a side business making robot toys. Their success in the toy market has been rather limited.

Their best known product is My Real Baby, manufactured by Hasbro around 1999-2000. It's basically a baby doll with Furby-type software. Rated "Worst Idea of the Year" by the Alliance for Childhood. It's not even that original; Baby Think it Over, the anti-teen-pregnancy doll from hell ("requires real care on the part of the student, including feeding, burping, rocking, and changing diapers"), has been around for years, but at a price well above the toy level.

This whole direction is way too much like Eliza. Much of the AI field, having failed at tasks that actually require doing something successfully without human assistance, now seems to be focused more on faking it. You've all seen Ask Jeeves, and obnoxious "virtual customer support reps". Those are pathetic.

There's some good work going on, but this isn't it.

Re:Rod Brooks and the Toy Factory

[mizhi]

It's not that AI has failed, it's just that it has failed in the "HAL 9000" sense of the term and had unexpected success in other areas. Most modern algorithms and stuff that have to deal with a certain amount of instability in the world owe their success to thousands of AI researchers bashing their heads against a wall in the vain hopes of creating artificial life but coming up with all sorts of interesting observations and ideas to deal with specific problems. For example, slashdot ran a story on this little machine. You can sure as shit bet that the software involved in causing that thing to walk has at least a little bit of algorithms and ideas incorporated into it from research in AI.

Re:Rod Brooks and the Toy Factory

[Animats]

Sounds a bit like science in general; ie: looking for something else and stumbling onto a whole slew of other things.

AI research may have been most useful as a money acquisition scheme for expensive research, back when computers were expensive and rare. It's hard to believe today, but AI research used to dominate computer science at the major schools, which were MIT, CMU, and Stanford.

Spinoffs from AI research include time-sharing, EMACS, electronic mail, document processing, and parts of the ARPANET. But most of those came from the support people, not the AI researchers. Stallman, for example, worked at the MIT AI Lab for years, but he's not an AI person. AI lab support staffs made all the useful stuff work so the AI researchers could get their email, do graphics, move files around, and do all the basic computer stuff we now accept as normal.

The AI researchers helped justify using multimillion dollar machines for trivial stuff. Without them, nobody would have dared do, say, e-mail. When the hardware got cheaper, we knew how to do lots of useful, but not previously cost-effective, basic tasks. That may be the greatest contribution of the AI community.

Another "gee whiz" Media Labs story; move along

[Dr. Spork]

Am I the only person who thinks the Media Labs are just a bunch of frauds who do mickey-mouse science in order to generate stories in the New Scientist and Scientific American Frontiers (which in turn bring more funding)? I mean, this fish thing isn't their lamest project (these are the same guys who pay big bucks to research dog training over the internet), but if this same research were coming out of some considerably less prominent school than MIT, people would look at it and say "hey kids, stop fucking around already and make something important."

Toy companies!

[peter303]

A hit toy can return hundreds of millions in profit. The trend is toward more "active" toys. A.I.-bots are likely candidates whether they are on the screen or mechanical.

Re:What effect will this really have on the sea..?

[Xaoswolf]

Not to mention the fact that a herd(?) of dolphins will beat the living snot out of anything that attacks them if they can't readilly swim away.