The moment an artificial #intelligence accidentally wiped out #humanity...
The following section is taken from the #essay by #BruceSchneier about The Coming AI Hackers:
Imagine a robotic vacuum assigned the task of cleaning up any mess it sees. It might disable its vision so that it can’t see any messes, or cover messes up with opaque materials so it doesn’t see them.72 In 2018, an entrepreneurial—or perhaps just bored—programmer wanted his robot vacuum to stop bumping into furniture. He trained an #AI by rewarding it for not hitting the bumper sensors.73 Instead of learning not to bump into things, the AI learned to drive the vacuum backwards because there are no bumper sensors on the back of the device.
source: https://www.belfercenter.org/publication/coming-ai-hackers
This means, in its #consequence, that AI would unhesitatingly wipe out humanity if that would bring it closer to the programmed #reward. We could not even call this AI #evil, because it has no emotions at all and knows no human #morality. Ultimately, it would not be the AI's #mistake either, but in fact the fault of unclean, incomplete programming. Now look at the quality of today's software with Trojans, viruses, backdoors, spam and all that crap. Does anyone seriously believe that the #military and intelligence services do not use all the software possible? Against fully autonomous #combat #robots, one can only defend oneself with combat robots. Either with quantity or better AI. Once #Skynet is activated, the solution could be to wipe out humanity, because without humans there are no more conflicts on Earth and world #peace is, after all, the highest reward for such a combat #robot.
#software #problem #ethics #future #bug #hack #hacker #conflict #war #science #emotion #defence
