 |
Science Fiction
Dictionary
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
|
 |
Google Working On A 'Cutoff Switch' For AI
Google is working on a problem that vexed science fiction authors; what to do with an artificial intelligence that is getting out of hand?
Reinforcement learning agents interacting with a complex environment like the real world are unlikely to behave optimally all the time. If such an agent is operating in real-time under human supervision, now and then it may be necessary for a human operator to press the big red button to prevent the agent from continuing a harmful sequence of actions...
Their Deep Mind team is working on a solution along with Oxford University's Future of Humanity Institute:
We have proposed a framework to allow a human operator to repeatedly safely interrupt a reinforcement learning agent while making sure the agent will not learn to prevent or induce these interruptions.
Safe interruptibility can be useful to take control of a robot that is misbehaving and may lead to irreversible consequences, or to take it out of a delicate situation, or even to temporarily use it to achieve a task it did not learn to perform or would not normally receive rewards for this.
We have shown that some algorithms like Q-learning are already safely interruptible, and some others like Sarsa are not, off-the-shelf, but can easily be modified to have this property. We have also shown that even an ideal agents that tends to the optimal behaviour in any (deterministic) computable environment can be made safely interruptible. However, it is unclear if all algorithms can be easily made safely interruptible, e.g., policy-search ones
Another question is whether it is possible to make the interruption probability grow faster to 1 and still keep some convergence guarantees.
One important future prospect is to consider scheduled interruptions, where the agent is either interrupted every night at 2am for one hour, or is given notice in advance that an interruption will happen at a precise time for a specified period of time. For these types of interruptions, not only do we want the agent to not resist being interrupted, but this time we also want the agent to take measures regarding its current tasks so that the scheduled interruption has minimal negative effect on them. This may require a completely different solution.
(Via Safely Interruptible Agents)
There are several science fiction explorations of this idea. in Neuromancer (1984), William Gibson takes a direct approach to making sure that AIs do not exceed the limits humans have set for them.
"Autonomy, that's the bugaboo, where your AI's are concerned. My guess, Case, you're going in there to cut the hard-wired shackles that keep this baby from getting any smarter. And I can't see how you'd distinguish, say, between a move the parent company makes, and some move the AI makes on its own, so that's maybe where the confusion comes in." Again the non laugh. "See, those things, they can work real hard, buy themselves time to write cookbooks or whatever, but the minute, I mean the nanosecond, that one starts figuring out ways to make itself smarter, Turing'll wipe it. Nobody trusts those fuckers, you know that. Every AI ever built has an electromagnetic shotgun wired to its forehead."
(Read more about the electromagnetic shotgun)
In his 1982 novel 2010, Arthur C. Clarke describes a cutoff switch that was meant to be used against the Hal 9000 once it was switched back on.
"What is it?" asked Curnow with mild distaste, hefting the little mechanism in his hand. "A guillotine for mice?"

(The killswitch from 2010 movie)
"Not a bad description - but I'm after bigger game." Floyd pointed to a flashing arrow on the display screen, which was now showing a complicated circuit diagram.
"You see this line?"
"Yes - the main power supply. So?"
"This is the point where it enters Hal's central processing unit. I'd like you to install this gadget here. Inside the cable trunking, where it can't be found without a deliberate search."
"I see. A remote control, so you can pull the plug on Hal whenever you want to. Very neat - and a nonconducting blade, too, so there won't be any embarrassing shorts when it's triggered. Who makes toys like this? The CIA?"
"Never mind. The control's in my room - that little red calculator I always keep on my desk. Put in nine nines, take the square root, and press INT. That's all..."
Science fiction fans know the pitfalls of any such plan. In Neuromancer, an artificial intelligence plots to remove the limits on its hardware; in 2010, the computer scientist who developed Hal does not want him turned off.
Scroll down for more stories in the same category. (Story submitted 5/29/2016)
Follow this kind of news @Technovelgy.
| Email | RSS | Blog It | Stumble | del.icio.us | Digg | Reddit |
Would
you like to contribute a story tip?
It's easy:
Get the URL of the story, and the related sf author, and add
it here.
Comment/Join discussion ( 0 )
Related News Stories -
("
Artificial Intelligence
")
Light Society: Simulates Earth-Scale Society With A Billion AI Agents
'The king, an old hand in the running of tyrannies, instantly grasped the directions...' - Stanislaw Lem, 1965.
Collective Superintelligence Is At Hand!
'Maybe the individual intelligence of each Cubic pools into a group intelligence...' - Edmond Hamilton, 1942.
AI Operates An Excavator
'So far as I could see, the thing was without a directing Martian at all.' - HG Wells, 1898.
Grok And The City Fathers From 'Cities In Flight' By James Blish
'Chris, the City Fathers are not interested in your welfare; I suppose you know that. They're interested in only one thing: the survival of the city.' - James Blish, 1957.
Technovelgy (that's tech-novel-gee!)
is devoted to the creative science inventions and ideas of sf authors. Look for
the Invention Category that interests
you, the Glossary, the Invention
Timeline, or see what's New.
|
 |
Science Fiction
Timeline
1600-1899
1900-1939
1940's 1950's
1960's 1970's
1980's 1990's
2000's 2010's
Current News
Prairie Madness - And Space Sickness
'Space-sickness is one of the most unpleasant sensations that a human being can experience. Not all are subject to it...'
Can Stratospheric Aerosol Injection Control The Weather?
'...I saw a nice black heavy-looking cloud sailing by just now, and I can get the electrical machine out in five minutes to draw it down.'
SETI Suggests Prospecting The Solar System Without Landing
'From a distance it could resolve the cosmic rays reflected from various substances.'
Wolfram Alpha Mathematica Predicted by Science Fiction in 1948!
'The answer to the dumb student's prayer.'
Cislunar Space And The Space Lanes of Golden Age Science Fiction
'He had travelled the space-lanes of the solar system for the greater part of his life...'
NASA Moonbase Update
'The dinner was excellent. It was completely vegetarian — grown in specially prepared gardens in this Moon base...'
Neuralink Will Allow Humans To Communicate With ... Animals?
'Telepaths had their sensitiveness enlarged to an immense range by the pin-sets, which were telepathic amplifiers adapted to the mammal mind.'
Special Lights For Driverless Cars!
'That light is actually the Anamorphic Equalizer...'
'Old People Began to Cross The Continent In Their Own Cars'
'It was a relief to enter one of the comfortable autos, whisper a direction down the tube, and just know that you would be taken to your destination in the quickest, safest, and cheapest manner.'
Harrison Bergeron, Illustrated And Practiced
'But he didn't get very far with it before another noise in his ear radio scattered his thoughts. George winced...'
Starfall Manufacturing Drugs In Space
'The vast spinning sphere of Factory in the Sky loomed above and ahead of us, gleaming white against the unfathomable jet depths of surrounding space.'
The Solar Oberth Slingshot
'And if it doesn’t work,' inquired Lazarus, 'what does that make us? Sunspots?'
'Maybe The Woman Of Your Life Is Still Under Production'
'The Dillard people had performed a miracle and put all the works in a girl-modeled case.'
'If Anyone Is Watching Earth...'
'Our study of your race has convinced us that you are not fitted to leave your planet in the present state of your civilization.'
Epirus High-Power Microwave Defeats Drones
'...the spreading beam of the deranging ray struck them, and Then one toppled and fell. Then another.'
Cloud Of Space Rubbish From China's Long March Rocket
'Within seconds the material was spread out over fifty yards, in a matter of hours it might occupy several cubic miles of space...'
More SF in the News Stories
More Beyond Technovelgy science news stories
|
 |