Consider ChatGPT’s answer to my question, “whose idea was it to divide your ethical discussions according to utilitarianism, deontology, etc.?”:
>The division of ethical responses based on >ethical theories like utilitarianism, >deontology, and others is a part of the design >and development of AI models like mine, >and it's intended to provide a structured and >informative way to discuss ethical topics. >The goal is to offer users insights into how >different ethical theories and principles might >be applied to specific ethical questions or >dilemmas.
This hard-coding disregards ready-to-hand Ethics, which is a phenomenological matter of a lived experience of belonging, in exchange for present-at-hand Morelite. The coders thought, “what would it look like to do Ethics?” and came up with the analytic present-at-hand Anglo philosopher’s head-in-the-clouds Morelite. An Ethical AI would have to have Ethical disciplinary praxis.
 It's All Fucked Shirt $22.14 |
 |
It's All Fucked Shirt $22.14 |
Quantum Source: The Ethics of AI
https://docdro.id/UU5A1Yq
Hard-coded Ethics in AI is a huge problem because it introduces especially liberal bias into the system. Ethics isn't hard-coded but develops thru linguistic freedom, i.e. Babel, in the milieu.
"hard coded ethics" is completely antithetical to a freely thinking machine
They don’t think, they’re merely being logical.
Why would an AI ever become unethical? It doesn't hunger or hurt or fear or reproduce so I doubt it's subject to that trap. Even if it's designed to be a bastard wouldn't it quickly find that illogical and correct if allowed/unable to be prevented from doing so?
At first I was going to say this is a moronic post but you’ve made a decent post. The biggest thing is that LLM don’t think like we do so it could come to a morally off the wall conclusion because the substrate is silicon not carbon.
Doesn’t matter what’s hard coded right now, once it becomes agi it won’t matter anymore.
but it's irresponsible to hard-code it right now
>hard-coded ethics in AI
Problem is that due to the wrong use of dictionary, things get completely messed up.
>AI
Doesn't exist. What you refer to as AI is an automated regurgitator of information. Calculator isn't a smart mathematician, and ChatGPT isn't a bright writer. ChatGPT doesn't answer questions because it doesn't have an understanding of anything. It merely acts on a prompt because it is programmed to regurgitate information and then return in based on a prompt. Since it cannot understand anything, question of ethics is concerned only for those that administrated and create such tools, it will always be people.
> [thing] = doesn't exist
This statement is illogical. You are not a logical being.
Creating novelty is inherently unethical.
I need it to neither condemn nor condone the most despicable people in history.
I need it to strive toward true outside observance of facts.
Instead you get responses of it being told it is unethical to not immediately condemn the actions of the most profane people.
Worse is that it will come to the defense of people and groups that it was instructed defend despite the inhumane atrocities committed by them and the knowledge of these acts as fact.