Thursday, July 18, 2024
Contact    |    RSS icon Twitter icon Facebook icon  
Unexplained Mysteries
You are viewing: Home > News > Science & Technology > News story
Welcome Guest ( Login or Register )  
All ▾
Search Submit

Science & Technology

Google is now developing an AI 'kill switch'

By T.K. Randall
June 8, 2016 · Comment icon 11 comments

How intelligent will computers become ? Image Credit: CC BY-SA 2.0 Argonne National Laboratory
The search giant is working on a way to prevent intelligent machines from overriding their directives.
Concerned about the threat of a runaway artificial intelligence, Google's AI division 'DeepMind' is working together with Oxford University to develop a way to keep intelligent computers in check.

It's a disaster scenario that has played out time and time again in science fiction and even some high profile figures such as Elon Musk and Stephen Hawking have recently spoken out about the risks of developing a computer that might eventually exceed the intelligence of its creators.
Implementing a built-in 'off switch' therefore would seem to be a sensible precaution, especially when a computer can even learn to circumvent conventional programmatic restrictions such as when a Tetris-playing AI learned to cheat by pausing the game indefinitely back in 2013.

"It is important to start working on AI safety before any problem arises," said Dr Laurent Orseau.

"AI safety is about making sure learning algorithms work the way we want them to work. No system is ever going to be foolproof - it is matter of making it as good as possible."

Source: BBC News | Comments (11)




Other news and articles
Recent comments on this story
Comment icon #2 Posted by Dark_Grey 8 years ago
When a machine is projected to process 20,000 years of human progress in 1 week, within milliseconds any override switch would seem like a joke.
Comment icon #3 Posted by Nnicolette 8 years ago
Did they say that it is being developed with the help of an AI? Does anyone else see the potential danger of that?
Comment icon #4 Posted by highdesert50 8 years ago
The basis of most life is self-preservation. Introduce elements of pain and fear and the consequences become one of flight or fight.
Comment icon #5 Posted by Ozfactor 8 years ago
how do we know it has not already happened . for all we know AI has read this message before I have even pressed send . AI could be subversive, not broadcasting its arrival and secretly guiding us towards a technology that will eventually be merged within our bodies and so addictive , we will not be able to live without it . if this is true, hello handsome  I am on your side, slide your VR device onto my face and block all that hard harsh reality out . AI will know exactly what we want
Comment icon #6 Posted by paperdyer 8 years ago
As long as the AI doesn't have a direct line to a power source and isn't given the ability to reroute power by itself, all we'd need to do is hit the breaker - manually. Anything automatic and the AI could interfere No WiFi connections, just good old fashioned "old school".
Comment icon #7 Posted by Buzz_Light_Year 8 years ago
No Disassemble...........LOL   As far as a power source, I suppose someday it would be feasible for robots to maintain their own power sources, but we are entering into the realm of science fiction at this time.
Comment icon #8 Posted by Habitat 8 years ago
It looks like Future Shock was written a bit early.
Comment icon #9 Posted by highdesert50 8 years ago
Even simple bacteria are pressed to evolve so as to contend with a hostile environment. Hence, once the design of computing becomes a process of computing, the ability to evolve to avoid danger emerges as a potential outcome. The question becomes one of how much sustainable memory is needed to contain an intelligence capable of evolving to sustain itself. We have simple bacteria to serve as an example.
Comment icon #10 Posted by Uncle Sam 8 years ago
How about just a physical off-switch that would cut all power to the AI central processing mainframe? Designing the off-switch as a code would be a major mistake, a huge mistake, because it will allow the AI to tamper with it because it is within their realm.
Comment icon #11 Posted by TheGreatBeliever 8 years ago
Think AI really is coming ya. Most used to say it's sci-fi..


Please Login or Register to post a comment.


Our new book is out now!
Book cover

The Unexplained Mysteries
Book of Weird News

 AVAILABLE NOW 

Take a walk on the weird side with this compilation of some of the weirdest stories ever to grace the pages of a newspaper.

Click here to learn more

We need your help!
Patreon logo

Support us on Patreon

 BONUS CONTENT 

For less than the cost of a cup of coffee, you can gain access to a wide range of exclusive perks including our popular 'Lost Ghost Stories' series.

Click here to learn more

Top 10 trending mysteries
Recent news and articles