Losing control of artificial intelligence (AI) is the biggest concern around the computer science, the Technology Secretary has said.
Michelle Donelan said a Terminator-style scenario was a “potential area” where AI development could lead but “there are several stages before that”.
I have questions…
What if the rogue AIs mocked out genitals mercilessly before killing us? I think that would be worse.
And how many stages? Is anyone keeping an eye on developments? Because they have drones hunting and killing humans now. Robot dogs with guns. Who get to pull the plug on research that is too risky?
It’s not rogue AI, it’s dumb prediction engines getting sold as intelligence.
Guessing once a cover-up gets exposed where AI systems ended up causing friendly fire or a mass casualty from hitting an unknown edge case.
We are still a long way from any sort of AGI from dystopian science fiction. But “AI” can still be very unintelligent but still be potentially very dangerous, and the sort hyperbolic doom mongering that keeps coming up in this discussions is distracting from that.