If you want to feel paranoid about AI, then Google?s deep mind is proof positive of what will happen when the robots take control.
Apparently, DeepMind's AI recently acted aggressively when threatened in a computer game and threw its toys out of the pram.
DeepMind boffins performed several tests by having its most complex AI play a series of a games with a version of itself.
In the first game, two AI agents, one red and one blue, scramble to see who can collect the most apples, or green squares.
Each AI has the option of firing off a long laser beam to stun the other AI, giving one player ample time to collect more precious green apples.
It took no time for both AI to start zapping each other relentlessly in the name of green apples.
The researchers say that this shows that AIs can learn aggressive policies in environments that combine a scarcity of resources with the possibility of costly action.
The boffins could change details in the game that would push the AI into being more or less likely to zap their apple-gathering partner.
This lowered the frequency at which apples spawned and also upped the stun time for each AI's laser, tweaks that resulted in a "highly aggressive" game.
While we are big fans of anything that is highly aggressive about destroying Apple, it does mean that the technology could be swiftly deployed to aggressively destroy humanity.
Reply With Quote
Bookmarks