3 ms·
what will ground an AI in human ethics? What grounds humans in "human ethics?" Is it thoughtful rational decision making or whims of the powerful at the time?
by AndrewKemendo 9y ago
what will ground an AI in human ethics?
What grounds humans in "human ethics?" Is it thoughtful rational decision making or whims of the powerful at the time?
And which human ethics are you grounding in? Aristotelian, Cynic, Stoic, Theological?
There seems to be a common idea that humans all share some ethical framework or thread that is similar. I don't think that holds. In fact I don't know if there is any single ethic that you could argue every single human holds similarly. Maybe you could argue, will to survive and reproduce, but even that is highly variable.
The idea that we should or could build some kind of ethical homogeneity into our engineered intelligent systems is absolutely silly.
- RobertoG 9y agoThere are some characteristics, shaped by evolution, that are typically human. For instance, humans are social. That doesn't mean that you can derive a formal system of ethics that is "true", but there are clearly some trends. Trends that an AI, that have not evolved as us, don't need to share. I don't think is a sterile debate.
- AndrewKemendo 9y agoFor instance, humans are social As are honeybees. I fail to see how this (or any) feature is distinguishing when describing how one should build an ethical framework into an engineered system. If however there was a more nuanced pro-social univerally attributable aspect that could be described and modeled, then I think that would be a helpful addition. I don't think is a sterile debate. It's anything but sterile. However what is sterile is the idea that there exists a desirable ethical system for AGI that is based on human ethics.
- RobertoG 9y ago>>"I fail to see how this (or any) feature is distinguishing when describing how one should build an ethical framework into an engineered system. " My point is that human ethics is grounded in human values and those are grounded in our evolutionary history. As we share an evolutionary history it would be not strange that there are a lot of values that we share. Maybe they are difficult to see because they appear to us as obviously true, but this is not going to be the case with machines. Presumably, we want machines that care about the same things that us.
- AndrewKemendo 9y agoMaybe they are difficult to see because they appear to us as obviously true I argue they are difficult to see because they don't exist. For example "Don't Kill" would be the obvious values that we share - until it's not a value we share. Maybe it's a value we ought to share, but as the dozens of examples in philosophy show (trolley problem etc...) in fact it's not something that we can hard code. Ok, so maybe it's "it's ok to kill in the following circumstances..." in which case a brief history of humanity would show massive changes in what those circumstances are over the last 10,000 years - so that's a bad candidate for a "value" set as well. Presumably, we want machines that care about the same things that us. This is the going argument and has been since the concept of "friendliness" in AI came about. Trouble is, we can't seem to pin down the things we as a human race collectively "care about" to the granularity that would be required to write machine instruction around it. I'm arguing there aren't any collective values that everyone would agree upon which would inform such a discussion.
- bobthechef 9y agoIt seems you grossly misunderstand ethics by portraying it as some kind of arbitrary system (or "framework" per your word choice, a word I've come to despise) imposed on people, or some kind of code voluntarily adopted but ultimately without ground. People may disagree about what is good, or what the correct moral principles are, but that does not mean there isn't a proper object of study and that that object is objective and universal. Cultures can be wrong. People can be wrong. People can be ignorant. Their moral intuitions and their ethical judgments can be deformed and at odds with the truth. In any case, moral skepticism and moral relativism are not only worthless, but incoherent and untenable and nothing anyone of any philosophical sophistication holds to. It is the quintessential freshman philosophy student's position, a position born out of ignorance and teenage rebelliousness rather than the careful exercise of reason.
- AndrewKemendo 9y agoPray tell, what "objective and universal" object are you referring to?
- JoeAltmaier 9y agoThou Shalt Not Kill is a pretty good start
- DmenshunlAnlsis 9y agoYeah, right. The problem is that if you think militaries don’t have plans for AI you’re nuts. Blanket prohibitions against killing and automated killing machines are incompatible. The civilian sector can be as moral as it wants, yet ironically it will be the AI most able to kill that will be least restrained from killing.
- tacon 9y agoI know that the very first time I watched the MIT robot cheetah, I instantly imagined the version that is surely coming: that robot with a gun mounted on it, hunting down a human[0]. [0] https://www.youtube.com/watch?v=_luhn7TLfWU https://www.youtube.com/watch?v=_luhn7TLfWU
- AndrewKemendo 9y agoThis is a common example used, to which the common question is "Is there absolutely no case in which killing is a justified and moral position?" Even then you must specify further. "Don't kill" but only humans or all things? If another human is attacking you with intent to kill and you have no other options? How about in a case like the trolley problem [1] where you are forced into a challenge of scale? These are all valid problems with designing current engineering systems, not just AI systems. It's possible to philosophically "bite the bullet"[2] and say "No, it's never justified or moral" however that doesn't pass the test of practicality. So, something as seemingly obvious as "Don't kill" doesn't have firm boundaries which we can create a universal set of mathematically implementable, and acceptable, rules for. [1]https://en.wikipedia.org/wiki/Trolley_problem https://en.wikipedia.org/wiki/Trolley_problem [2]https://en.wikipedia.org/wiki/Bite_the_bullet#In_philosophy https://en.wikipedia.org/wiki/Bite_the_bullet#In_philosophy
- JoeAltmaier 8y agoIs there no spectrum between absolutes? Anyway the AI will have no concept of the notion, unless its part of the calculation. And Utilitarian notions are notoriously repugnant.