Should religious thinkers help shape AI?

Should we try to tune incomprehensibly complex, black box software - created by machine-learning algorithms - to be more "moral" when simulating a human conversational partner by adding religious texts to the training set? It may be a good way to make it better at simulating a religious human but, as with many religious people, its decision process may then appear to be dictated by scripture, rather than guided by an internal moral compass. So religious texts may well make the chatbot appear less moral - just as they do with many human beings.

I don't think it is good that religious thinkers help train AI, since their forma mentis has nothing to do with what AI is and specifically how it works out. Religious thinkers are moral bent pundits, they are prone to limit free research and to encircle free thought. Give back to Caesar what is Caesar’s and to God what is God’s...

Por supuesto que sí. Si se quiere llegar a la AGI si, sin duda. Argumentación: Si se entiende la AGI como una propiedad emergente de la humanidad, pretender excluir cualquier aspecto de ella (la humanidad) es sesgar y, por tanto, limitarla. Sobretodo si el aspecto excluido hace referencia a las más altas cualidades del ser humano, como la religión (de religare: “unir lo que no está unido”). Aclaración: Una cosa son las máximas potencialidades y virtudes de un concepto, y otra muy diferente el uso polarizado de esas virtudes con el objeto de imponer mi visión, o destruir y denigrar lo que no estoy de acuerdo. Como analogía… ¿Tendría sentido excluir la ciencia y a los cientificos del desarrollo en IA porque inventaron la bomba atómica y se utilizó? Segundo: Evitar el sesgo de la religión concreta, entender cada una de ellas como expresiones del “hecho religioso humano” en ese aspecto la pregunta planteada es limitada en cuanto incluye a los “pensadores religiosos” ¿Y los místicos? ¿Y los …? Tercero: Conviene diferenciar y no hacer chocar los diferentes usos de la IA. Una cosa es el análisis masivo de datos para reconocimiento de patrones y otra muy diferente el control (en su aspecto de gestión y ayuda) de la humanidad. Aspecto al que todo el mundo quiere llegar primero pero pocos se atreven a confesar. El debate planetario, profundo y público hace tiempo debería haber comenzado. Vamos tarde y nos estamos jugando la supervivencia de la especie (o mucho sufrimiento) por arañar unas cuantas miguitas más de poder al otro

Absolutely! It is the only amendment.

I want to add 1 more point to my earlier comments.

In my opinion AI currently must stand for Artificial Intellect and not Intelligence.

Intelligence will always stand and work for the Greater good.

I am an atheist. Religion or non-religion is the right of every individual, his personal choice. Therefore, every human being has the right to participate. Science alone cannot be the determinant of our future.

Is religious tradition a useful source of moral wisdom for AI ?

If AI was to classify religion as a source of moral 'wisdom'—using ‘all’ of its purported intelligence—it would/should simply describe religious beliefs, ‘religions scripture’ as simple parables, dogma, fairy tales, fiction. Books, fairy tails, written words, that are 'mans' self described divine words that allows many to escape reality as well as responsibility. Fiction written by the self ordained 'experts'. Should one say demagogues, on the speculative subject. Not unlike any author of any fictitious writing/command where the masses are called upon to defend—to suffer and to kill for the ‘chosen’ one—in the name of whichever fictitious hominid that the so called experts introduce to them who will willingly accept as 'gods'—gods that demand total and unquestioning allegiance, faithful to the/their end.

If AI was to describe religion it MUST describe religions true purpose of controlling the/a narrative of death and destruction; from and before the crusades, including ‘rules/laws’ that promote and advance gender violence, manifest destiny, doctrine of discovery, slavery, greed, fear and condemnation of them others; The list goes on and on! And All must be included!

It must also include the notion of dominion over the lands air water, and seas. Domination that allows the righteous to kill all living things, including that that sustains life itself, our mother earth! ...Of course in the name of god! The history of all the pain, death and destruction thought history must be meticulously documented and brought out front and center!

So: Should the science of AI include religion? Only if the nature of religion is described to be what it is!!!

Should chatbot values be shaped through secular, technical processes instead?

Understanding, teaching, demonstrating, caring for, including explaining, 'technically', how the environment, our mother earth, is in fact crucially, essential to life itself; Your life. It's life.

Water is Life!! This is a must...Before and as we kill ourselves:

Programing Computers to sounding alarms, 'worldwide', alerting the entire world at the first hint that one or some are imperiling ALL!!

One should question: Doing unto others as you would have them do unto you...There are millions and millions of people who love to torture and be tortured, kill and be killed.

All religions and their proponents are based on blind ignorant belief. Further, most are corrupt. The tech industry is as bad or worse. Please use only secular scholars, thinkers and philosophers.

i think that there should be a combination of theologians and concerned citizens of a diverse compostion. It's like trying to bring up a child. What ethics will that child be taught, what principles of life will the child be brought up on. Then after a certain age that child has to be set free to navigate the world. Make mistakes, continue to evolve as the world evolves. This is the challenge with AI.

The one does not exclude the other.

Please, just use modern secular technical processes, if any.

AI and chatbot values should be shaped according to secular ethics, not according to religious tradition neither trough technical processes.

Secular ethics derives ethical principles and values from human reason, empathy, and experience as well as universal human rights rather than religious doctrine or divine revelation.

Hmmm...from a purely feminist POV I'm going to have to say "hellllls no!'. I agree with the commenter who mentioned the Stackhouse 3-dimensional, comprehensive ethical framework approach, the multi-layered consideration of right v. wrong, good v. evil, just(fitting) v. base seems like a much sturdier tripod than the theologian's one-legged stool approach. Plus, there's already FAR too much dissension in those ranks for comfort...Christianity alone houses a plethora of contradictions these days, politically and otherwise. Color me Staunchly Concerned.

There should be no Chatbots. There is no need! People should talk to people. This is what is wrong wth AI.

The question of whether religious tradition should or should not be included in AI chatbots is really the wrong question. Most religions recognize God even though mankind, as a whole, is unable to identify who or what is God. Should a chatbot make the presumption that there is or isn't a God? A chatbot is really limited in its' attempt to create answers regarding anything due to the limited physical nature of a chatbot. A chatbot cannot include the full scope of life, let alone the simply question regarding God's impact or not. Chatbots need to keep it simply.

I think this discussion should focus on the biggest dangers of AI which include dissemination of false information, making AI weapons in the form of robots and drones, and ultimately making the majority of the human population unnecessary for the wealthy people controlling AI.

I think it is laudable that Anthropic is trying to make their AI behave ethically, and getting input from philosophers, ethicists and theologians of ALL faiths will probably do more good than harm. If Claude is asked a "morally charged" question, perhaps it could specify answers based on the questioner's stated religious beliefs. Claude the Imam, Claude the Pope, Claude the Rabbi, Claude the Atheist, Claude the Christian Fundamentalist, Claude the Wiccan, etc., etc., etc.

But really, the only moral/religious tenet needed is The Golden Rule, viz., do undo others as you would have them do unto you. If religions would restrict themselves to that rule, much tragedy will be avoided. Unfortunately, too many religions believe that their way is the ONLY way and that all others are false and should be condemned if not eliminated.

Based on their actions, none of religions have right to claim any hold on morality. Not any more than socialist parties or dictators. History did happen, wars, persecutions and tortures were and still are committed in the name of these very same religions. Thanks, no thanks. Monogamy is not a virtue, it is a choice, unlike the sectual orientation that is not a choice. Morality is not a value on its own, it is a mean to an end - productive and safe existence of individuals and societies. Contracts we have with each other to be able to trust others that they too will contribute to the society when they commit to do so.

We need a much broader discussion for the AI, and because its influence, the societies governance. Let start at the basics. Is lying moral? Is corruption moral? Unless we do not agree on these questions we are in trouble. Again, it is not about Good or Evil, it is about survival. Liars make us do dangerous things and corruption results in incompetency of deciders, that again, results in dangerous actions. I would like the Humanity to survive, despite our failings. We must not destroy the life on Earth and, to the best of our abilities, educate the next generations so they have a fighting chance. If some kind of Morality Codex will help us with that, let's do it. But I am sorry, but I have no confidence into belief systems that for thousands of years burn people alive for pointing at the ship disappearing beyond the horizon.

Our knowledge is not perfect and is ever changing, but it is our best tool to try to predict the outcomes of our actions and correct them when beneficial to all of us. United we stand, divided we fall.

Yes, of course. All thinkers should be part of the discussion and to exclude some because of other views is neither smart nor science. Should their opinions be weigted more than atheists -- no.

I am fairly sure that the general public is becoming aware of the possible dangers that AI could present but aren't able/willing/inclined to do anything about it. The public is also becoming skeptical that politicians are even interested in putting guardrails on the AI industries. If religious thinkers can try to influence the public and, through them, the governments to take these dangers seriously, that would be worthwhile.

If we humans are lucky, in the future there will be no religions. They are all fictions. Morality is independent of religion. Ethical codes do not need religion to be viable.

I am a committed Christian, but not an evangelical. I believe in a freely available open church but not in pursuing converts or harassing folks with whom I disagree. Therefore, I would rather generative AI not favor any particular religious tradition, including non-christian traditions, but include all traditions. In other words, all religious traditions are appropriate sources of moral wisdom as well as secular traditions.

I too am encouraged that "AI companies are trying to decide how chatbots should respond to morally charged questions . . ."

Without religious/theological input an indispensable element of several thousand years human thought, experience and accumulated wisdom would be excluded. We must not function as if current thinkers have all the answers, superior to all who have gone before.

How about sticking with the golden rule and forget about all the religious trappings. Imagine an AI that decides it is God-like and acts accordingly.

I'd as soon they consulted ethicists, rather than theologians. It seems these companies (it they're serious) want to build their AI's to be "moral." Which is a good thing, but one can (as the Hippocratic Oath says) "do no harm" outside of the colorings of religious thought.

That said, I do find it admirable that Anthropic is attempting to create a "moral, ethical" AI platform. I wish them well. It will surely help when Anthropic and other companies actually build a real artificial intelligence, rather than the predictive LLMs currently in use.