Bishop Thomas Dowd

A parishioner trusted an AI over their own bishop. The AI was wrong.

I get an email, Bishop, are you sure we're allowed to do that? I thought we could only do this. I said, well, where did you get that idea? Well, I asked magisterium AI. I said, aha. So in my job, I'm being questioned as to my competence based on what an AI said.

A bishop on AI and prayer: a machine cannot pray on your behalf.

a machine cannot pray on your behalf. Like if, if I was going to do a rosary with 150 Hail Marys and I said, well, I don't have time to say 150. I'll just get my computer to say it. No, that's not going to work.

Even if only 1% turn to AI partners, that's 80 million people.

is everybody going to have an AI boyfriend or girlfriend? No, I don't think so. But if just 1% of the population is experiencing that kind of existential loneliness, well, that's 80 million people. That's a lot of people.

The next divide may not be wealth or intelligence. It may be emotional capacity.

And so you'll wind up with a stratification of human beings, not based on wealth, not based on intelligence, based on the emotional capacity to enter into relationship with others. What kind of society will emerge from that? I have no idea, but AI could certainly accelerate that big time.

The real nightmare isn't a psychopathic AI. It's psychopathic people running it.

Not that the AI becomes psychopathic, but that the people behind it are able to use it for just for their own interests with no empathy for anyone else, seeing the human beings interacting with it as ones and zeros, effectively, just as extensions of a system.

At six or seven, I decided I'd rather live an uncomfortable truth than a comfortable lie.

I realized I would rather live an uncomfortable truth than a comfortable lie. And it seems so self-evident to me. It was a nice moment, but it seemed self-evident. And I must have been very young, like six or seven or eight.

A priest's blog raised the money for a school library in one week.

We had all the money in one week, one week. Because people, what I realized is because people knew who I was and because of the content, it was building a sense of connection. There was also a sense of trust.

The problem with chatbots: an answer can seem true and not be.

Had I trusted the answer, I would have made errors. And so this is the problem. There's a level of response that can seem true, but which isn't.

His AI gave a wrong answer, and the fix was blocked by copyright.

And they said, oh, we can't input that because it's under copyright. So in other words, you can only train your AI on older garbage instead of the newer great stuff. Terrific. What do we need this for?

An AI may know what a rose smells like and still never have smelled one.

Angels don't have bodies. And so they know what a rose smells like, but they can never actually have the experience of smelling a rose. You know, they could talk about the smell of a rose, but they can't do so from a point of view of experience.

We don't let airplane makers self-regulate. Why would AI be different?

The profit motive is one that has never really been trusted for pure self-regulation in any circumstance. There's a reason why we have regulatory agencies for the construction of airplanes, right? There's a reason why we have regulatory agencies for the crafting of vaccines or the approval of medications.

AI can feed our appetite to consume. It can't do the accomplishing for us.

I don't think AGI can ever replace the irascible appetite because you want to be the one playing football, not your robot. You know, you want to be the one getting a diploma. If you're in school and you're saying, the reason I want to get a degree is so I can get a good job, why not cheat?

You can't smell incense online, and communion won't come by DoorDash.

You can watch it online, but you can't smell the incense online. You know, and there's, you can't receive communion on your tongue online. And we're never sending it by DoorDash.

If no atom has free will, where does ours come from? A bishop's argument against machine minds.

From my point of view, the critical question is free will itself. None of the atoms of our brain individually possess free will. And so collections of them wouldn't possess it either. That doesn't mean they don't possess randomness.