Has anyone attempted to build moral principles into an AI model?
Thank you for your reply and I think you are in the same headspace as I am on this topic. I will check out the links you posted.
A side note: what is the difference between ethics and morals?
Hi you havent replied to the comments people made, youve replied to your post. People wont get a notification that you replied.
Ok, thank you, I’m still getting used to the buttons on here
deleted by creator
Huge literature on this, hard to know where to start.
Search arXiv or a similar site.
Great thanks, I will check it out.
All of the info on AI development seems to be hush hush so I am very curious what it is going on with developing AI.
I remember a couple years ago there were calls to cease AI development with concerns about world safety and then the commercial media outlets sort of stopped talking about it.
Ethics, not morals, but yes, this is a core part of “alignment” or making sure the machine wants the same thing we do. It turns out alignment is really hard because ethics is really hard. The classic AI doomsday story is based on an AI that took utilitarianism as the highest end goal (the best way to save humanity is to destroy it); that’s an ethical framework used to justify genocide.
So shorter answer: “Yes, but ethics is hard”. I really like Robert Miles’ videos on this topic; here’s one to get you started: https://www.youtube.com/watch?v=ZeecOKBus3Q
And here’s another very related one for after: https://www.youtube.com/watch?v=hEUO6pjwFOo
Thank you for your reply and I think you are in the same headspace as I am on this topic. I will check out the links you posted. A side note: what is the difference between ethics and morals?
My pleasure!
Morality has a religious basis and ultimately comes down to “God’s will is how things should be, so how close is it to God’s will?”
Meanwhile, ethics is secular and tries to answer questions of “right” and “wrong” in the absence of an omnipotent power.
They already did.
‘Absurdly woke’: Google’s AI chatbot spits out ‘diverse’ images of Founding Fathers, popes, Vikings https://tech.yahoo.com/ai/articles/absurdly-woke-google-ai-chatbot-185153814.html
Why Google’s ‘woke’ AI problem won’t be an easy fix https://www.bbc.com/news/technology-68412620