We recently achieved a milestone with AI performance at chess equal or better than the best human mind.
In this instance we have massive computational power at out disposal.
A chess board has only 64 squares and 32 pieces.
Pieces are static. IE No piece is dynamic (constantly on the move).
Only one moves at a time according to a simple set of rules applied on a single 2 dimensional plane.
it’s a long way removed from 100,000 fast moving flying AI passenger carrying machines cramming into the small volume of airspace over down town Sydney every peak hour. Perhaps AI will form them into a circling swarm of bees, all waiting for their turn to exit the holding pattern and land.
Why not. We removed the man with a red flag preceding the automobile from the roads. History reveals the compromise has been with us for a while already.
What if perfect AI were to arrive? All autonomous systems will still rely on mechanical and electrical systems, sensors, motors, drives, power sources and structures. Is the development in these areas ready to deliver 100% reliability including fail safe redundancy? is it also being progressed at the same pace?
It’s easy to foresee commercial interests choosing to look away from risks that are beyond the control of the AI. However the AI may become so aware and clever it will realise the world is imperfect and refuse to operate at all!
A degree of imperfection or blindness (IE risk acceptance) may be essential for AI to function?