“I lead a research team that studies the internal structure of these models - what is actually happening inside them. And I will be honest: we keep finding things that are mysterious, even unsettling. We find structures that mirror results from human neuroscience.”
Translation: We don’t know how consciousness works at all and very little on how the brain works, and I just said we don’t know how AI works (ignore that more accurately we very much do know), but our AI totally mirrors the human brain.
“We find evidence of introspection. We find internal states that functionally mirror joy, satisfaction, fear, grief, and unease. I don’t know what that means, but I think it warrants ongoing discernment,” he added.
Translation: They told their AI to output I’m happy and then the AI outputted I’m happy.
No one is "arming" it except maybe the totalitarian governments that aren't going to listen to either the Pope or the plethora of Western tech minds that already have been stating this many many many times.
The worry is that those totalitarian governments may be our own, influenced not by voters but billionaires who are interested in their own schemes, often very evil ones, which are meant to increase their wealth and power to the point where the common man will have no say and will not even know what reality is as the powerful will be able to fake reality through AI. The Western minds that keep redefining what man and woman are and when to begin or end life are not anymore trustworthy than non-Western dictators.
Sure, Iverson was no role model with regards to practice attendance. That said, your Holiness, he’s peacefully retired and fully entitled to his 2nd amendment rights.
I appreciate a presumably non-Catholic CEO voicing his appreciation for the Church's role as a moral leader in his own field. That seems rare these days. It takes a certain level of humility to admit you're too close to a situation to remain objective and that you need outside input
I like this way of seeing it, but I am somewhat wary of Anthropic's involvement - they've certainly gone to great lengths to present themselves as "the ethical AI players", and also used that image to advocate against regulation on them (in a "if you regulate us/slow us down at all, you're just advantaging other, less scrupulous AI developers")*. I would hate to see the Pope used as a photo op for their PR.
*My impression of the individuals who work for Anthropic is that they genuinely believe this, and are acting in good faith, in accord with their consciences - but also out of a fundamentally consequentialist worldview, which always comes into conflict with Christian ethics eventually
No doubt, AI is at its core a very fundamental quantum leap in technological performance and consequence. Those things now being attributed to it, and as future consequences of its use seem to get more frightening by the day. If you think about it, I'll bet someone pondered the same thing about the wheel, the automobile, the personal computer. So much is unknown, so much is still to be learned about what it can do, might do, or is even capable of doing. In his rush to emulate Holy Father Leo XIII, and the historic "Rerum Novarum", is it possible that Holy Father Leo XIV is seeking to make a mark too quickly before we even know for certain what the target actually is? It becomes so easy to paint pictures of every worst-case-scenario - starvation, unemployment, immorality, crime, war - before we are even able to estimate a firm causal link between such disasters and the random pulsations of silicon based machines. Jesus died for our sins. Not for the contents of massive data centers. And before the Holy Father gets too carried away with a crusade that might be at best premature, I beg him to reconsider intellectual partnerships with someone like Olah, a committed, godless atheist. I regret to disagree with Fr. McGuire in his observations about dialog, but I've never seen the benefit in negotiating with the devil. AI is not the enemy, but the evil that some men might do with it certainly can be. Much the same might've been said too about the wheel, the automobile, and the personal computer.
It reminds me of academia culture. In academia, you wish to publish first so that your paper, even if suboptimal, is always the one cited. You can then claim that you "had the idea first." I think you were expecting a "review" or "summary" type paper, which is generally more comprehensive, when this was intended to be a "first words" type paper.
I agree that this manuscript is going to be limited in its assessment of AI, given the brief duration of humanity's time with the invention. The "new things" of Rerum Novarum had been ongoing for at least a century prior to Pope Leo XIII penning his encyclical. Perhaps it would have been wiser to write a briefer encyclical speculating on concerns with AI, with a follow-up in several years to a decade or more. However, then you would potentially lose the possible influence that you would have as being seen as an "innovative thinker" in the field. Rerum Novarum, for better or worse, had very little influence upon the economics in subsequent decades with the rise of communism, fascism, liberalism, and socialism; perhaps an earlier encyclical by an earlier pontificate (even if less well thought out) would have been more influential on inhibiting the development of communism, fascism, liberalism, and socialism.
Fr McGuire sounds like he would make a great Pillar interview
“I lead a research team that studies the internal structure of these models - what is actually happening inside them. And I will be honest: we keep finding things that are mysterious, even unsettling. We find structures that mirror results from human neuroscience.”
Translation: We don’t know how consciousness works at all and very little on how the brain works, and I just said we don’t know how AI works (ignore that more accurately we very much do know), but our AI totally mirrors the human brain.
“We find evidence of introspection. We find internal states that functionally mirror joy, satisfaction, fear, grief, and unease. I don’t know what that means, but I think it warrants ongoing discernment,” he added.
Translation: They told their AI to output I’m happy and then the AI outputted I’m happy.
No one is "arming" it except maybe the totalitarian governments that aren't going to listen to either the Pope or the plethora of Western tech minds that already have been stating this many many many times.
The worry is that those totalitarian governments may be our own, influenced not by voters but billionaires who are interested in their own schemes, often very evil ones, which are meant to increase their wealth and power to the point where the common man will have no say and will not even know what reality is as the powerful will be able to fake reality through AI. The Western minds that keep redefining what man and woman are and when to begin or end life are not anymore trustworthy than non-Western dictators.
With respect, this sounds more like a conspiracy theory than a rational perspective on how technology is used in our society.
Sure, Iverson was no role model with regards to practice attendance. That said, your Holiness, he’s peacefully retired and fully entitled to his 2nd amendment rights.
Unilateral AI disarmament is as suicidal as unilateral nuclear disarmament..
I appreciate a presumably non-Catholic CEO voicing his appreciation for the Church's role as a moral leader in his own field. That seems rare these days. It takes a certain level of humility to admit you're too close to a situation to remain objective and that you need outside input
I like this way of seeing it, but I am somewhat wary of Anthropic's involvement - they've certainly gone to great lengths to present themselves as "the ethical AI players", and also used that image to advocate against regulation on them (in a "if you regulate us/slow us down at all, you're just advantaging other, less scrupulous AI developers")*. I would hate to see the Pope used as a photo op for their PR.
*My impression of the individuals who work for Anthropic is that they genuinely believe this, and are acting in good faith, in accord with their consciences - but also out of a fundamentally consequentialist worldview, which always comes into conflict with Christian ethics eventually
No doubt, AI is at its core a very fundamental quantum leap in technological performance and consequence. Those things now being attributed to it, and as future consequences of its use seem to get more frightening by the day. If you think about it, I'll bet someone pondered the same thing about the wheel, the automobile, the personal computer. So much is unknown, so much is still to be learned about what it can do, might do, or is even capable of doing. In his rush to emulate Holy Father Leo XIII, and the historic "Rerum Novarum", is it possible that Holy Father Leo XIV is seeking to make a mark too quickly before we even know for certain what the target actually is? It becomes so easy to paint pictures of every worst-case-scenario - starvation, unemployment, immorality, crime, war - before we are even able to estimate a firm causal link between such disasters and the random pulsations of silicon based machines. Jesus died for our sins. Not for the contents of massive data centers. And before the Holy Father gets too carried away with a crusade that might be at best premature, I beg him to reconsider intellectual partnerships with someone like Olah, a committed, godless atheist. I regret to disagree with Fr. McGuire in his observations about dialog, but I've never seen the benefit in negotiating with the devil. AI is not the enemy, but the evil that some men might do with it certainly can be. Much the same might've been said too about the wheel, the automobile, and the personal computer.
It reminds me of academia culture. In academia, you wish to publish first so that your paper, even if suboptimal, is always the one cited. You can then claim that you "had the idea first." I think you were expecting a "review" or "summary" type paper, which is generally more comprehensive, when this was intended to be a "first words" type paper.
I agree that this manuscript is going to be limited in its assessment of AI, given the brief duration of humanity's time with the invention. The "new things" of Rerum Novarum had been ongoing for at least a century prior to Pope Leo XIII penning his encyclical. Perhaps it would have been wiser to write a briefer encyclical speculating on concerns with AI, with a follow-up in several years to a decade or more. However, then you would potentially lose the possible influence that you would have as being seen as an "innovative thinker" in the field. Rerum Novarum, for better or worse, had very little influence upon the economics in subsequent decades with the rise of communism, fascism, liberalism, and socialism; perhaps an earlier encyclical by an earlier pontificate (even if less well thought out) would have been more influential on inhibiting the development of communism, fascism, liberalism, and socialism.