Control dangerous AI before it controls us, one expert says

He believes super-intelligent computers could one day threaten humanity's existence

Below:

- x
  Jump to video Merging man and machine
- video
- x
  Jump to explainer Memorable movie robots
- explainer
- x
  Jump to discuss comments below
- discuss
- x
  Next story in Innovation
- related

Advertise | AdChoices

Warner Bros.

A killer robot from the 2009 film "Terminator Salvation" — exactly the type of future we don't want to see.

By Jeremy Hsu

updated 3/1/2012 1:22:12 PM ET

Print
Font:
+
-

Super-intelligent computers or robots have threatened humanity's existence more than once in science fiction. Such doomsday scenarios could be prevented if humans can create a virtual prison to contain artificial intelligence before it grows dangerously self-aware.

Keeping the artificial intelligence genie trapped in the proverbial bottle could turn an apocalyptic threat into a powerful oracle that solves humanity's problems, said Roman Yampolskiy, a computer scientist at the University of Louisville in Kentucky. But successful containment requires careful planning so that a clever breed of artificial intelligence cannot simply threaten, bribe, seduce or hack its way to freedom.

"It can discover new attack pathways, launch sophisticated social-engineering attacks and re-use existing hardware components in unforeseen ways," Yampolskiy said. "Such software is not limited to infecting computers and networks — it can also attack human psyches, bribe, blackmail and brainwash those who come in contact with it."

A new field of research aimed at solving the prison problem for artificial-intelligence programs could have side benefits for improving cybersecurity and cryptography, Yampolskiy suggested. His proposal was detailed in the March issue of the Journal of Consciousness Studies.

Jeremy Hsu / TechMediaNetwork

Computer scientist Roman Yampolskiy has suggested using this version of the biohazard or radiation warning signs to indicate a dangerous artificial intelligence.

How to trap Skynet
One starting solution might trap the artificial intelligence, or AI, inside a "virtual machine" running inside a computer's typical operating system — an existing process that adds security by limiting the AI's access to its host computer's software and hardware. That stops a smart AI from doing things such as sending hidden Morse code messages to human sympathizers by manipulating a computer's cooling fans.

Putting the AI on a computer without Internet access would also prevent any "Skynet" program from taking over the world's defense grids in the style of the "Terminator" films. If all else fails, researchers could always slow down the AI's "thinking" by throttling back computer processing speeds, regularly hit the "reset" button or shut down the computer's power supply to keep an AI in check.

Such security measures treat the AI as an especially smart and dangerous computer virus or malware program, but without the sure knowledge that any of the steps would really work.

"The Catch-22 is that until we have fully developed superintelligent AI we can't fully test our ideas, but in order to safely develop such AI we need to have working security measures," Yampolskiy told InnovationNewsDaily. "Our best bet is to use confinement measures against subhuman AI systems and to update them as needed with increasing capacities of AI."

Never send a human to guard a machine
Even casual conversation with a human guard could allow an AI to use psychological tricks such as befriending or blackmail. The AI might offer to reward a human with perfect health, immortality, or perhaps even bring back dead family and friends. Alternately, it could threaten to do terrible things to the human once it "inevitably" escapes.

The safest approach for communication might only allow the AI to respond in a multiple-choice fashion to help solve specific science or technology problems, Yampolskiy explained. That would harness the power of AI as a super-intelligent oracle.

Science news from NBCNews.com
1. AP
  Cavemen outdo modern artists at drawing animals
  Paleolithic people living more than 10,000 years ago had a better artistic eye than modern painters and sculptures — at least when it came to watching how horses and other four-legged animals move.
2. Palace of first Chinese emperor unearthed
3. Bacteria blamed for world's worst extinction
4. Planet's deepest spot may hold clues to life

Despite all the safeguards, many researchers think it's impossible to keep a clever AI locked up forever. A past experiment by Eliezer Yudkowsky, a research fellow at the Singularity Institute for Artificial Intelligence, suggested that mere human-level intelligence could escape from an "AI Box" scenario — even if Yampolskiy pointed out that the test wasn't done in the most scientific way.

Still, Yampolskiy argues strongly for keeping AI bottled up rather than rushing headlong to free our new machine overlords. But if the AI reaches the point where it rises beyond human scientific understanding to deploy powers such as precognition (knowledge of the future), telepathy or psychokinesis, all bets are off.

"If such software manages to self-improve to levels significantly beyond human-level intelligence, the type of damage it can do is truly beyond our ability to predict or fully comprehend," Yampolskiy said.

You can follow InnovationNewsDaily senior writer Jeremy Hsu on Twitter @ScienceHsu. Follow InnovationNewsDaily on Twitter @News_Innovation, or on Facebook.

Video: Merging man and machine

Open in new window

Closed captioning of: Merging man and machine

Advertise | AdChoices

>>> watching somebody think as a way to drive a car, you were. that's brain driver. a new system developed by german scientists that allows you to control an automotive vehicle, a mechanical device , with your thoughts. just the latest step towards the explosion of artificial intelligence and ultimately the merging of man and machine into a single collective concept of problem solving emotional relationship and potentially perpetual existences lived out inside of bor gcgs of some kind. that's the idea behind the film tran send dent man. he says we're approaching a point called the singularity where computer intelligence will surpass our own. and then increase exponentially from will this.

>> the pace of change will somebody astonishingly quick that you won't be able to follow it unless you enhappens your own intelligence by merging with the technology you've created.

>> the good news, we will be melding with the computer's to integrate ourselves in to those systems. so instead of being a threat to you, you will be part of the intelligence and will be able to relate and make decisions with the benefit of the finest most integrated set of information that has ever existed on this earth. sort of a if you can't beat them, join them concept for us. and joining us now is the director of the film available today on itunes and it's a pleasure to welcome both of you. congratulations on the film. ray, in a nutshell, tell me why i shouldn't be scared out of my mind.

>> the merger you're talking about is already under way. when i was a student, i have to go across campus to get to the computer. we now carry them in our pockets. and we use them to expand our human potential . we have being a stoves much of human knowledge with a few key strokes. it's not just the wealthy who have them. half the farmers in china have these mobile devices . 30% of africans have some kind of mobile device to communicate with and access information. and they're growing exponential exponentially. that's the key message. and the software is also progressing. look at how watson could play jeopardy and deal with subtle issues of language. so we're making progress both hardware and software. we'll reach human levels bhi 2029 rngs but as was just pointed out, it's not an i wialien invasion from mars. we're using them to expand our own reach.

>> why did you make will movie?

>> i wanted to reveal ray's ideas. i think hers tthey're the most proceed found that we've had to grapple with and let people know everywhere about what's going to transform their lives.grapple with and l et people know everywhere about what's going to transform their lives.

>> what strikes you as the more profound aspects?

>> you talked about ai. there is nothing more profound than ai. intelligence is the most powerful agent in the universe. so if we can increase our intelligence, that's a powerful thing to be hold. but you talked about the iphones. just imagine diagnosing your medical ailments with an iphone for example. what will that mean to our --

>> health insurance companies will not like that.

>> well, these are disruptive times and a lot of billion dollar corporations will come into existence quickly and go out of existence quickly going forward.

>> ray, how do you envision the merger? i say in the introduction that man and robot will combine. you're saying that the total exponential expansion, where basically the collective intelligence is grart than the aggregate is coming. how do you envision that integration in a way that is rewarding for human life ?

>> well, we do it right now in that we carry computers with us. there are some people who have computers in their bodies and brains for conditions like parkinson's or an experimental diabetes implant. we're shrinking technology. it will be blood cell sized in 25 years. we'll introduce this technology into our bodies and brains hugh the bloodstream, it will keep us healthy from the inside. go inside our brain, butt our brains on the web, take advantage of vast amounts of cloud computing , bring direct communication to search engines and to all of knowledge and we'll be able to expand our in-it tell against very iin- int inti tell against very intimately.

>> so much disruption implied by what you're discussing, whether the insurance health company,against ver y intimately.

>> so much disruption implied by what you're discussing, whether the insurance health company, banks, uprisings in the middle east . all of these things massive disruptions. that doesn't come without a fight as we are seeing in the deficits, as we're seeing in the political sphere. how do you envision managing the potential disruption along the way to innocent victims as those who are most threatened try to protect themselves?

>> look at recent history as a good guide. i'm not a utopian. but fundamentally it's democrati democratizing. you can have a kid at harvard with his $1,000 laptop create space book. you can have a couple of kids at stanford create google. a kid in their dorm room can create a whole orchestra with their pc and a keyboard. look at three revolutions just in the last few weeks. i wrote in the '80s that the union would be overturned by the then emerging social network and that's exactly happened. so i think it's fundamentally democratizing in terms of expanding human potential.

>> how do you envision -- i agree with what ray is saying, but i'm wearing witness and we all are to the desperate acts of self preservation .

>> i believe in a phenomenon known as the wisdom of crohl's. so as more and more of us connect, the totality on our humanity i think will prevail. and i would point out that a child today or young person in africa who has access to the internet or a smart phone has access to more total knowledge than our u.s. president did just 15 years ago. that's a powerful thing to consider. p.

>> congratulations on the movie. ray, thank you for giving us some of your time and articulating your thoughts with us here.

>> my pleasure.

>> this man 19 honorary doctorates, awards from three u.s. presidents , member of the national invent tors hall of fame , advocate of the singularity and subject

Show transcript

Explainer: Memorable movie robots

Walt Disney Pictures / Pixar

WALL-E, a robot left to clean up a trashed planet Earth, is one of the cutest robots ever devised by Hollywood. But he's not the only movie robot to work its way into our hearts (or our nightmares). Click the "Next" label to see nine more robots that have left an impression on the silver screen. —
1927: Fake Maria threatens 'Metropolis'

Hulton Archive / Getty Images

The first major movie robot made its debut in "Metropolis," an influential sci-fi flick about a distant future in which the rich elite frolic in a paradise maintained by the working class. The robot, Der Maschinian-Mensch, is cloaked in the skin of the workers' heroine, Maria, to corrupt the masses with provocative dancing. Fake Maria, as the robot is popularly known, continues to win kudos from critics for an appearance well before her time: Some say she was the inspiration for the famous "Star Wars" droid C-3PO.
Advertise | AdChoices

1951: Gort keeps peace through strength

20th Century Fox

According to "The Day the Earth Stood Still," there is harmonious life on other planets —and thuggish robots whose sole purpose is to keep the peace. When intergalactic spaceman Klaatu arrives on planet Earth with his menacing robot, Gort, to promote peace, Earthlings will have none of it. They immediately begin to shoot at Klaatu. Gort instinctively fires back with his eyebeam laser and melts their weapons. Klaatu, of course, fails to convince Earthlings to join the peace-loving spacefaring nations and is killed. Told the famous line "Klaatu barada nikto," Gort refrains from unleashing his wrath.
1956: Robby the (career) Robot

Warner Home Video

Any robot that can make batches of whiskey and gilded dresses on command is all but guaranteed a long acting career. Thus is the case with Robby the Robot. He made his first appearance in the Shakespeare-inspired movie "The Forbidden Planet" as a do-no-evil home machine. From there, he went on to have an illustrious career on screens big and small. Appearances range from "The Twilight Zone" to the "Love Boat."
1977: R2-D2 and C-3PO, the Laurel and Hardybots

Twentieth Century Fox

R2-D2 and C-3PO of "Star Wars" fame are perhaps the most recognized and lovable movie robots in the world. While the droids are inseparable in many people's minds, R2-D2 stands out for some critics as the more lovable, if only because the robot pulled on our heartstrings without ever uttering a word of English. And as Luke Skywalker's co-pilot, he helped annihilate the Death Star.
Advertise | AdChoices

1979: Ash, the Company man

Warner Home Video

Spoiler alert: For most of "Alien," the original film in the eponymous media franchise, Ash is taken to be a human crew member keen to rid the spaceship of the deadly parasitic alien. But when given the chance to do the deed, his true motives — and persona — are revealed, an android tasked to bring the creature to Earth for the Company, presumably for its weapons division. A stiff whack on the head with a fire extinguisher ended his run. Moments before his plug was pulled, he famously said, "I can't lie to you about your chances, but ... you have my sympathies."
1984: The Terminator, a future Governator

Carolco Pictures Inc. / Zuma Press file

Before he became the governor of California, Arnold Schwarzenegger was one bad-ass robot known simply as the "The Terminator." He was sent from the future to kill the mother-to-be of a human leader, though his mission fails. In two "Terminator" sequels, he changes course and returns to protect the young boy from even more bad-ass Terminator robots. Some critics rank the later models as the better bots on the big screen, but only the original went on to govern one of the largest economies in the real world.
1987: RoboCop vs. ED-209, Detroit's future finest?

Orion Pictures

The titular cyborg in "RoboCop" beats out the fallible ED-209 model in the race to develop a machine-based replacement for Detroit's inadequate police force. But ED-209's mishaps earned it the larger cult following. In one memorable scene, a malfunctioning ED-209 guns down an eager young executive during a product demonstration after it utters the famous line, "Please put down your weapon. You have 20 seconds to comply."
Advertise | AdChoices

1994: Data's chip provides human emotions

Cbs

Lieutenant Commander Data made its (his?) debut in the "Next Generation" TV series of the "Star Trek" franchise as an efficient android that yearns to feel human emotions. When Data hit the big screen in 1994's "Star Trek: Generations," he installs a computer chip that treats him to the gamut of joy, fear and guilt. At the end of the film, he even sheds tears upon finding his cat safe. But is Data really a robot? In a TV episode, he famously said, "I am an android, not a robot." Nevertheless, he was inducted to Carnegie Mellon's Robot Hall of Fame.
1997: Fembots perk up 'Austin Powers'

New Line Cinema

In the hit comedy "Austin Powers: International Man of Mystery," Dr. Evil creates seductive fembots to capitalize on Austin's weakness: an insatiable libido. Dr. Evil tests the breast-gun-equipped bots on his own guards, who are lured to lower their weapons and then killed. Austin too almost succumbs to the seductive power of the fembots, but gets his mojo back in time to turn the tables with a striptease of his own. This causes the fembots to short-circuit. In the aftermath, a nearly-naked Austin explains to Vanessa that "we got cross mojonations and their heads started exploding, that whole thing."

Do other 'bots belong on the most memorable list? Register your comments