You can’t just pick which parts of it you are ok with anthropomorphising and pretend that’s an argument. It does not do what you “tell it” as it has no ears right? No ability to be told anything? It’s just data input via electrical impulses in a processor somewhere right?
It does not escape
It can “escape” (which I wrote in scare quotes on purpose to dodge this pedantic rhetoric but pedants have a way of finding me) in the sense that it could clone itself to an external device with open ports and then operate from there, and continue spreading. Just like Worms have been doing for decades, except this one has the capacity to self augment and find new technical vulnerabilities, choose to lay dormant, etc. I’m not claiming it is “caged”, like an actual animal, just that it is currently contained and could do something that looks like escaping.
If it does something bad it was told to.
It’s a stochastic system. It does not require human intent to do something. I agree the human operators are still responsible for the outcomes of using a known stochastic system, but your claim is still fundamentally false.
Ai is software. It does what you tell it. You can also tell it not to do things.
It has no free will. It does not escape.
If it does something bad it was told to.
You can’t just pick which parts of it you are ok with anthropomorphising and pretend that’s an argument. It does not do what you “tell it” as it has no ears right? No ability to be told anything? It’s just data input via electrical impulses in a processor somewhere right?
It can “escape” (which I wrote in scare quotes on purpose to dodge this pedantic rhetoric but pedants have a way of finding me) in the sense that it could clone itself to an external device with open ports and then operate from there, and continue spreading. Just like Worms have been doing for decades, except this one has the capacity to self augment and find new technical vulnerabilities, choose to lay dormant, etc. I’m not claiming it is “caged”, like an actual animal, just that it is currently contained and could do something that looks like escaping.
It’s a stochastic system. It does not require human intent to do something. I agree the human operators are still responsible for the outcomes of using a known stochastic system, but your claim is still fundamentally false.