Mr. Bruce Schneier’s article “What do you do with criminal AI?” was recently posted here and, wishing to keep discussion in house, I will comment here. To summarize an excellent question posed in the above titled piece by Bruce Schneier — What can we do to punish autonomous AI crime?
I am going to ultra simplify this;
An AI that is truly autonomous and sentient is a product of its programming. This is true regardless of the programming’s source — internal or external. Hence just as a we are — (for the most part and with an agreement that I am avoiding centuries of philosophical study and argument) — sentient beings, a non-biologic sentient AI exists as an individual through their mind or the approximation of same. That approximation int he sentient AI non-biologic is programming
So, if there is moral and lawful authority to do so, the AI is reprogrammed.
I also believe that even as sentient biologic creations become more comparable (via genetic alterations as well as bionics) to non-biologic sentient entities, there will still be the propensity to give dominion to the biologics. And if recent writing about the uprising of the machines has any validity, apparently the machines will have a similar bent.
So — if a non-biologic life form violates the law in a felonious manner and no biologic entity can be determined upon which to lay blame as the non-biologic’s creator or programmer, then the non-biologic is reprogrammed.
And I am sure my ridiculous oversimplification of the problem will be embraced by all and peaceful agreement will spread rapidly — or not.