Page 1 of 1

A question of ethics with AI

Posted: Mon Mar 12, 2018 10:40 pm
by Mythbhavd
In Solar Express, Marcel makes a discovery. At one point, Alayna admits to herself that Marcel truly made the discovery. At what point does an AI, especially one that is self learning and has been functioning as long as Marcel, count as independent enough to be credited with discovery or co-discovery? I'm currently rereading the story and don't recall how self-aware Marcel is (I don't believe it's mentioned?), but if "he" is, then there is an ethical problem if "he" is not credited with co-discovery.

Re: A question of ethics with AI

Posted: Tue Mar 13, 2018 12:58 pm
by lmodesitt
That's an excellent point, but it also hinges on the nature of "discovery." If automated photo-comparison software picks out an anomalous streak or point of light, is that the discovery, or is the intelligence [human or AI] that classifies it as a comet, asteroid, or whatever the discoverer? Is it the finding or the analysis that counts? I'm honestly not sure exactly where to draw the line. In the case in point, I think Marcel should get credit as co-discoverer [even if I didn't think that point through when I wrote the book].

L.E. Modesitt, Jr.,

Re: A question of ethics with AI

Posted: Tue Mar 13, 2018 7:59 pm
by Mythbhavd
I considered the question of programming, but Marcel seems to be more advanced than its original programming. One such instance is when Alayna asks a question about what she should do in a certain instance. Marcel's response is something along the lines of, "You seem to be asking me a question about morality, but I cannot make a moral decision for you." To me, that says that Marcel is actually able to consider moral implications of things, at least to a point. It was those thoughts that led me to the original question. If we were creating AI that could become self-aware, I think we'd still have the habit of thinking more of the AI as a thing to be used rather than giving it recognition of its personhood.

I bounced this idea off of a group of friends as we were talking on Marco Polo tonight. It has spurred an interesting discussion. One pointed out that man has shown himself to be oppressive. In his opinion, the default would be for man to claim credit even in those instances where AI was shown to be self-aware. He believes that society would set the norm and the fact that Marcel is not recognized as co-discoverer shows that society sees AI as servants, slaves, or just machines, even when self-aware.

It seems to me that you've touched on this in one of your novels, but for the life of me, I cannot remember which.

Re: A question of ethics with AI

Posted: Fri Apr 27, 2018 6:24 pm
by Lucent
On programming, I work in development and part of that includes writing analytics software. When a piece of software that I designed discovers some new trend or algorithms a more efficient way to process data, I don’t pat my processor on the panel and give it a raise. The tool, no matter how efficient should never be mistaken for senescence. In today’s world as technology advances to the point that the efficiency of AI and predictive algorithms surpasses biological capacity, it is important that we as a society do not allow ourselves to lose sight of this, the repercussions could be disastrous.

But I know what I’m reading next now!