on a layman level AI is kinda my hobby. That means I can't code it but understand to some degree the underlying mechanisms.
First off there NO "understanding" of the "reality" we live in (at all) - zero "consciousness". And there will never be. Not because our chips are not fast enough or any other technical limitation but inherently, its a limit like the speed of light. Makes no sense until you understand what consciousness entails.
How good AI is in answering your questions is primarily dependent on the data it is trained on. There will never be a 100% guaranteed flawless answer because that would require a real understanding of the greater world, including aspects that are even hidden from our "dayjob" mind.
In a limited environment with a limited task AI can do very well. Thats why it will replace some repetitive jobs that do not require any creativity or great flexibility beyond the parameters set by the task environment.
AI driven robots will work in factories and other places, maybe even homes and hospitals. If they are made human like they can certainly be "polite" and very useful but never really understanding what they are doing and why in a deeper sense. There is no "I" in the sense of self recognition. Of course it can simulate such "self" and gullible humans can and will fall for it (robot girlfriends come to mind).
There can be misguided robots with weapons - and they can be lethal. But that is the humans responsibility, not that of the AI.
