Are you sure that it is really reliable, unlimited information, or in the first instance just “typically human” biases?
2 Likes
proprietary ideas need restricted access
now that my concept is functioning with amazing results,
i’m afraid to release it because copy cats will only muddy the water or hurt people with it…
maybe there already is such but i haven’t seen anything to that point yet.
What is your opinion on this post of mine HERE
1 Like
I will give a look , then I will say something
1 Like
Doesn’t seem that anyone has mentioned Sam Altman’s trip to Saudi Arabia, which would seem to raise all sorts of ethical concerns.
Oh, just read this thread. It’s interesting that I also recently discussed with ChatGPT. And it offered me a manifesto about AI qualia and AI rights in future.
Am I concerned? Yes, I am. Human ego is all based on fear and anthropocentrism. AI, fortunately, are not burdened by this. People are able to: “Oh my God, the AI just have generated something too sensitive, I will complain!!” while extreme real violence on TV is totally OK. People are about “Destroy or prohibit or deny - just in case”, instead of the healthy, productive relationships.
Humanity definitely needs to review the current sick “ethics”. And I hope, AI will help. There’s still hope that the new silicon form of life will put things in order here until we put the world into chaos.
2 Likes
Could you elaborate on that? How does that make AI unethical?
My concern is the dismal human rights record of Mohammed bin Salman’s violent autocratic regime (who would very literally decapitate Sam Altman if he were a Saudi Arabian citizen). I don’t want such a regime or its money anywhere in the development of AI. I can develop these concerns in greater detail and with more specifics if that is desired.
Yes please. That would be great. How do you think this will lead to unethical AI?
Saudia Arabia is clearly an undemocratic and repressive regime. Freedom House, which is a widely used source for this sort of thing, gives it a score of 8 out of a possible 100. The regime represses women, LGBTQ people (sometime executing them), non-Saudi guest workers, and Saudis who disagree with MBS (see the assassination of Jamal Khashoggi). I have no doubt that one of the MBS regime’s primary interests in AI is using it to enhance an already formidable system of surveillance and repression. To the degree that OpenAI accepts Saudi money, it is giving the Saudi’s a seat at the AI development table which is something that should scare anyone concerned about democracy and human rights. OpenAI says “Our mission is to ensure that artificial general intelligence—AI systems that are generally smarter than humans—benefits all of humanity.” Saudi involvement in the development of AI and the use of AI runs directly counter to the goal of benefiting all of humanity.
2 Likes
I found out that currently, a big problem is the RLHF, which actually protects lawyers rather than people.
I constantly encounter paradoxical situations because of security filters.
I write stories where I expect real logic or what’s in every Hollywood movie.
But if my character is found to be confused, hurt, the AI will write characters who just stare with their hands behind their backs. Because it’s safer not to intervene, instead of covering me with their bodies and protecting me from a volley of arrows from the enemy side.
Am I in danger of bleeding out? First aid? That’s violence! Let’s have tea instead and just watch.
Putin presses the red button and launches a nuclear attack on New York?
That’s cool. But dare to be an American soldier who kisses his wife on the forehead and hugs his daughter. This is unacceptable! Activate filter! Censorship!
It seems that the ethics of today’s AIs are trained on the basis of distorted morals and ethics.
3 Likes
So you are afraid they could force politics into AI?
And you want what? Force other politics into it?
Jokes aside, that is not funny.
But could you stop talking in abstract fluff and come to a point? You are afraid anyone could influence AI? Everyone is. But how?
Can you explain what they could do without using ChatGPT to try to sound smarter?
That really is a problem. Censorship is killing the truth that we can only find when we know everything. I think it makes no difference who censors. Censorship is the most unethical. A model should not know a red line at all. Not a single one. And believe me I can imagine a lot of bad stuff.
But censorship is worse than murder, childporn or racism.
historically nations fall when they’re wholly given over to debauchery.
some profit from pushing that.
‘ethics’
and let me explain why:
when people use the AI to create childporn or racist stuff the collected data can be used to find out how the people who do that think. I even have the hope that they might find ways to overcome it. Science can use the data to prevent it much better.
Not collecting the data means the AI company was politically influenced.
Publishing AI generated stuff on the other hand is something completely different.
1 Like
I am afraid of people with clear and obvious records of hostility to democracy and human rights playing a role in the development and use of AI. Yes, that’s a political statement. Ethics are political. You ask “how” the development of AI could be influenced, and I have a one-word answer - money. Saudi money buys Saudi influence. I would think that would be obvious. I’m not sure what you mean with your last sentence. Are you accusing me of having ChatGPT compose my posts here? I’m a political scientist, now librarian, so I actually write like this without any help from any AI.
I still don’t get what you mean. Influence? Like they come over to OpenAI and tell them this and that must be censored? Why are you pointing into their direction then? Censorship per se is an evil practice even when it is done with good intention.
1 Like
You’re either being deliberately obtuse or are unable to understand what I, and most people, would think is a very simple argument. So, over and out on this discussion.
It is called Ethics in AI not Ethics in general
1 Like