Video: Humanity Meets AI Symposium: AI and Religion

How might AI influence our understanding of humanity, morality, and meaning-making? How will religious traditions and communities adapt to or shape the ethical frameworks guiding AI development? In what ways can religious perspectives contribute to the creation of a more equitable society amid the disruptions AI will bring to labor, governance, and social structures?

Religion and Public Life hosted a symposium that explored the profound ways in which artificial intelligence is reshaping human society, with a particular emphasis on the role of religion, the transformation of societal structures and capitalism, and strategies to reduce inequities as society responds to the sweeping changes brought about by AI.

The symposium equipped our audience with the tools and frameworks to critically engage with the ethical, spiritual, and cultural dimensions of this transformation. In an age increasingly influenced by AI, this symposium helped identify practical opportunities to shape a more humane and just society.

FULL TRANSCRIPT

SPEAKER 1: Harvard Divinity School.

SPEAKER 2: AI and religion. February 28, 2025.

CAROLINE: And so I am excited to introduce the fourth panel of today talking about AI and religion. And so I'm going to do bios with an updated bio for Jen because I took it off of your mobile website, which obviously has not included all the work you've done at the UN. So that's good. And so today, for the fourth panel, we have Brian K. Smith is the honorable David S. Nelson Professor and Associate Dean for Research at Boston College's Lynch School of Education and Human Development and a professor in the computer science department. He recently co-led a discussion at the Builders Artificial Intelligence forum in Vatican City on fostering AI expertise within the Catholic Church. Dr. Smith holds a PhD in learning sciences from the Northwestern University and a BS in computer science and engineering from UCLA.

Jen Louis, who we all already know, but leads the UNDP's efforts in AI and trust and safety and is an affiliate of Berkman Klein Center at Harvard. She's the founder of the Moral Innovation Lab, co-founder of Spiritual Care Partners, and previously led integrity and policy teams at Facebook, Meetup, and Google. Her Harvard Divinity School research explored how technology shapes moral futures. Jen speaks globally on tech governance, online risk, and moral innovation.

Thomas Arnold is a research associate in the Tufts Human Robot Interaction lab, where he studies AI ethics, moral norms in human robot interaction, and gender in robotic touch. He has developed ethics curricula at Tufts, co-organized the coding caring study for AI 100, and serves as program lead for AI and faith. He is completing a PhD at Harvard's committee on the study of religion, and is co-author of Ethics for Psychologists: A Casebook Approach.

And then finally, Greg Epstein is the humanist chaplain at Harvard and MIT, advising on ethical and existential issues. He's the author of Good Without God and Tech Agnostic, which was named 2024 best business book of the year by Porchlight. A former TechCrunch ethicist in residence, Greg has written for Time, MIT Technology Review, cnn.com, the Boston Globe, and the Washington Post on technology, ethics, and secular humanism. So please welcome them all. Thank you.

SPEAKER 3: Thank you, Caroline, for those bios. So together in general, we have been bringing together some humans. I want to note that there are no AIs speaking today. It has all been humans. And everyone has had-- we've had a very wide range of backgrounds. I would say that we are disproportionately academics-- so people with pretty close ties to academia and people who are doing research, people who have done degrees recently.

This panel narrows that focus even more, as all of our panelists have ties, not only to the work of academia, but also to religion or religious studies. In the case of Jen and Greg, those might be right. We'll talk about what the definition of religion is and what their religious beliefs and views are. I think you'll hear people's views evolving, which is interesting, not just within the context of technology, but also generally.

It is no surprise that at the Divinity School's AI symposium that we are having a panel on AI in religion. But I want to highlight, especially in the vein of religion and public life, how religion can permeate other, perhaps, all facets of life. And our panelists here today are bringing those rich experiences that include religion without necessarily having it operate in the foreground.

With that, I'd love to start with professor Smith. Thank you for coming over to us from Boston College, crossing the river. You recently attended the Builders Artificial Intelligence forum in the Vatican, as Caroline had mentioned. What was that experience like? What is the vibe in the Catholic Church around AI right now?

BRIAN K. SMITH: Hello! Hi, everyone. So thank you for that. Thank you for the invitation to be here. What a great group of people in this panel. Yeah, so recently, I was at-- I guess that was October-- at a meeting at the Pontifical Academy of Sciences in Vatican City. The Pontifical Academy is the oldest science academy in the world, as it turns out, which is interesting because it is in this place that few people can really get to.

That particular meeting was called the Builders AI forum. And what it did was it brought a group of software developers, who primarily work in Catholic Christian applications. So there were people like Long Beard, Exodus 90, Halo, people who are-- software companies that essentially really do have a mission of we are either trying to do evangelism, we're trying to do education, or we're trying to do spiritual support for people who are, I mean, broadly Christian. But certainly, some of those people were Catholic.

And so I guess the-- I should say, so there were the specific companies that were working on those things. There were a group of other people, obviously priests, and others, who were there representing, what are the church's views? Academics were there. And then there were other companies Microsoft, Hewlett-Packard, representation from large companies that are not necessarily tied to religious applications. But they, in fact, provide so much support that you have to speak to them to think about, where are we going to actually deploy our applications?

So largely, it was how do we build an ecosystem that would be supportive of, again, supporting the Catholic mission through software? Understanding that we need to collaborate with others. Some of those people being academics, some of us being industry, some of those being the church itself. And how can we set up the terms, the kind of relationships that we need to do to actually foster that?

So I heard earlier, as people were talking about this idea of, how do we work with Nathaniel from Andreessen Horowitz, for instance, talking about the partnering. There was a lot of discussion about that. How do we essentially partner as organizations to try and do things that make sense but that also fulfill the Catholic mission? Because we were in Vatican. So that could be other religions, as well, it just turns out that this was in Vatican City.

I think, broadly, that could apply to any religion. And I think people were trying to be very open about, what do we mean about spirituality, or any of these things? Because it was about support for prayer or other things like that. Educational outreach and then evangelism were basically the three themes of that.

So I will say this, what was it like? Well, the Vatican City, again, it's not easy to get into. When you're walking around this-- if you're walking around Vatican City and the Swiss army come up and they ask you questions, you reply kindly because they've got guns. They got guns, as I think I said, in the Vatican. I was like, hey, you guys have guns that make Texans-- yeah, I said it.

Anyway, so I was just like, America would be proud of these guns. We love our-- I mean, you guys got guns. But that was the broad sense of things. And we came away from that trying to, again, set up the first kinds of talks about, what would that mean to actually do these kind of collaborations. And second, there were more discussions about different units. Like how would we think about software? How would we think about deploying software, testing those things, finding out who are the audiences, things like that?

AI, of course, came up a lot because it was-- nowadays, almost anything when you say software, AI is going to be somehow-- if it's not there already, it's going to be the next step. And so those discussions about how the church wants to think about AI, and what does it mean in terms of the value system that it brings to the morals and the ethics that the church brings was also a huge part of that conversation.

SPEAKER 3: Super interesting. And I think this will come up a little bit more later. But also just for context, how did you get on the assignment to go on that trip?

BRIAN K. SMITH: Man, I have-- OK, I have no idea, I should say. Some of these things have been just this weird fluke. And I don't know if this, but on November 30, 2022, something happened. And I guess it's just-- I think for a lot of people, ChatGPT showed up. And as people started to realize, hey-- I've worked in artificial intelligence for a long time-- let's just say a long time. And yeah, I mean, John earlier said expert systems. And I was like, yep, I remember that kind of stuff.

And so I think it was partially just in nature of the work. And then as you start-- as people started asking questions, then it's word of mouth and people eventually-- because, I should actually say, it's Boston College, if you don't know, we are a Jesuit institution. So that's the other connection was that many of, I should say, the academics who there were generally-- they were either from Divinity schools, or they were like a Villanova, us, Santa Clara. So the schools that tend to be Catholic universities.

SPEAKER 3: Awesome, thank you so much. The next question is for Thomas. I know that you have done some work. You've done a lot of really interesting work. But some of your work has been focused on AI and faith. And I was wondering if you could just give us some context onto how that became an area of interest of yours. What's motivating that?

THOMAS ARNOLD: Thank you for that question because I've been waiting to report this biographical detail for decades.

[LAUGHTER]

SPEAKER 4: This is your moment.

THOMAS ARNOLD: This is the moment, Rockefeller Two. Rockefeller Two, if you don't know, it's about 150 yards that way. I attended a talk in the late 90s, where robotics and religion was the subject. I had nothing but disgust for the topic and the discussion. The climax of which was a Methodist minister, who shall go unnamed, to my right, saying I could absolutely imagine a robot baptizing someone. And I can absolutely imagine baptizing a robot.

That sent me up to my dorm room at Rockefeller. And it was a difficult two hours of trying to grapple emotionally. But why was that so wrong? And I say that not just to unload that upon you. But I think when it comes to connecting AI and faith, I think we do have to think about bodies and thinking about bodily reaction, thinking about disgust, thinking about taboo, and not just staying there.

I'm so grateful to be here decades later to have people and colleagues to air this out a bit, both because those intuitions can, I think, be on to something and because I think they can be misshapen and formed by certain cultural traditions, some of which we may want to hold at arm's length and rethink. So I was contacted a few years ago by some folks who had worked at Microsoft, who wanted to have a group who could talk about AI and religion. Because in their tech work they did not find-- it was often they couldn't talk about their own faith life. They couldn't talk about what AI might mean for faith.

And I threw some connection. I get the flakiness of it at random. Someone said religion and robots, there's this guy that is at Tufts Robot Laboratory, but he's from religious studies. And I don't know how he got there. And I don't know what he's doing. And I don't know what perversity might be involved. But he's in there.

And so they contacted me. And I've been working with a network of people, who are now widely international, thankfully, with multiple religious traditions around, OK, what does this mean? And I'll just end with one example of a research group I'm a part of, where we're looking at compassion. What is our ideal of compassion? And because of the way that people are starting to introduce the idea of AI, is compassionate AI possible?

You can think about my experience of baptism, those memories coming up. But we have thinkers from Buddhist tradition, Christian tradition, Islamic tradition, thinking about, OK, what is our ideal? What is the ultimate form of compassion? And let's get that on the table because that's going to inform what we end up judging as deficient or sufficient when it comes to the implementation of these technologies.

SPEAKER 3: The next question is for Greg. So for Greg, we may or may not use the term "religion" or "religious" to describe what he does. But factually, we can say that he is the humanist chaplain for Harvard and MIT. He is a graduate of this institution as well. So we have a lot of ties to HDS here today. And Greg, why did you become a chaplain?

GREG EPSTEIN: Let me think. I mean, there's why I became a chaplain then and there's why I'm still a chaplain. I'm not sure that the answers are the same. I became a chaplain because Harvard had a humanist chaplain. And it had, had one for about 30 years by the time I arrived here, almost 30 years, at least.

And I had spent my life, up to that point, fascinated by religion, but I wasn't religious. Fascinated by community, but I didn't really have a community to call my own until I discovered this thing that is called humanism, which is an organized movement. Sometimes, organized movement feels generous.

It's sort of a semi-organized movement of people who are trying to live-- I'll do the branding-- good without God. They're trying to live well, to emphasize meaning, purpose, ethics, community, compassion without ideas of a supernatural, et cetera.

And I loved the idea. And I wanted an interesting job, someplace to actually put that into action. And when I met the late first humanist chaplain here at Harvard, one of the first in the world, the first day I met him, I was like, that's what I'm doing with my life. You're going to be my career counselor. And indeed, he was for a few years until he hired me as his assistant. And then ultimately, I succeeded him.

But I'll just say that the reason I've stayed at chaplain-- because over the years, I've learned enough about myself and what was really motivating me at the time, some of which wasn't all that great, or all that pretty, to have bailed on this many times over. And what keeps me in the role, I think, is the idea that-- I believe in the word "humanism" is particularly apt because we're only human, because we're mortal, we're going to die. And we can't control that unless you believe that I can, which I think is a really dangerous belief.

And we do so much better when we pursue some kind of pro-human, loving, compassionate meaning in our lives. And it can be hard because there's so much meaninglessness, so much anti-human sentiment in the world that pushes back on any effort that we might make to actually affirm ourselves and one another and do good for ourselves and one another. And so it's a really-- I like to think of it now, it's like a rebellious choice.

It's like pushing back against the rock. The rock is eventually going to smoosh me, but it's very meaningful to push back on it and especially to find other good people to push back on it with. As Ken would say, we find real joy in that.

SPEAKER 3: Greg has written a book. When was it pub-- when was Tech Agnostic published?

GREG EPSTEIN: Tech Agnostic came out about 100 days ago or so-- 110, 120, something like that.

SPEAKER 3: Pretty fresh off the printing presses, which come up a lot, I guess. And the full title is Tech Agnostic: How Technology Became the World's Most Powerful Religion. Would it be fair to say that your interest in technology precedes your interest in humanism and chaplaincy? Has technology been a long standing interest for you?

GREG EPSTEIN: No, and why it desperately needs a reformation. No, that's key. The last 100 pages of the book are about-- I'm an atheist in my professional life and in my personal life, too. In the world of tech, I'm just an agnostic. Some people say they're building gods, literally so. I mean, for all intents and purposes, I can't stop them. And I can't prove that they're not right.

But I'd say, at the very least, I want to be quite agnostic and skeptical of any such claims. And to wonder, as I would with any religious claims about gods, what's in it for you that you're so confident that this is going to be a God? And so desperately needs a reformation. What I'm looking for is not to eliminate tech any more than I'm looking to eliminate religion. I'm looking to improve it. I'm looking for us to critique it so that we can build something much more humane.

SPEAKER 3: Jen, so your background is a lot like mine, in that you had worked in technology. And then you came here to Harvard Divinity School to do the same degree program that I'm currently doing. And when we were just talking earlier, you mentioned that you weren't religious when you came here. So why did you come to Harvard Divinity School?

JEN LOUIS: Yeah, I mentioned this in the past. I think I-- well, it's funny. My friends actually questioned the same thing. I would say that when I came here, my friends, especially in tech, were like, is this your come to Jesus moment? Are you OK? What happened to you at Facebook?

[LAUGHTER]

I was thinking, that's a good--

SPEAKER 3: Reasonable question.

JEN LOUIS: Good question, yeah. And I do think that, yeah, I was a little burned out. I look at the underbelly of the internet. That's so much of the work that I did. And I look at the ways that people are terrible to one another on the internet. And I felt like I was trying to help people like from harm and being victimized. And then I'm like oh, no, this problem keeps getting bigger. And I don't know if my contributions to it are the right contributions if the problem is getting worse.

And I mentioned this in the previous panel. And so I think I had these existential questions. And I thought maybe I should go just take some time. And I thought about pursuing philosophy. But it was friends of mine, Casper and Angie, who were here, and I had these big questions, and they were like, actually, maybe this is more about a practiced approach. And maybe you should really consider Divinity School to explore these questions.

And I'm like, oh, I don't know. Is this a place that I like? And so they offered me the opportunity to apply for special student, which is what I did. And that was fantastic. So those of you who don't know about this program, I highly advise that if you're not planning on becoming a religious scholar and you're not sure if you want to do a full degree program, that being a special student means you can come here, study, choose whatever you want. There's no requirements.

And it gave me some time. And the reason why I ended up going with ultimate is because I needed more time. I felt like I just skimmed the surface of what was intended to be a pretty transformative experience for me. This was such a different pedagogy and approach to my learning experience than what I had prior.

I've been so acculturated to a tech way of thinking about things, that it was such a delight to be challenged to feel into concepts, and to really question the ontologies with which I approach problem solving, and to also being in such a supportive and loving space that would allow me to be wrong about it. I think there's actually not a ton of room in the tech space, culturally, to admit when one is wrong. That's also part of the thinking of the reformation that's needed. Like, how do we get into shifting out of hubris and into humility and have that be a virtue? And I love being here because of that.

SPEAKER 3: And just before we move ahead, you mentioned that during your time here, you did-- or at least right now, you have become more religious and you're on somewhat of a religious journey. I was wondering if you could speak to that briefly and maybe a little bit on how it intersects with your past in technology and your thoughts on technology.

JEN LOUIS: Yeah, the crazy thing is I thought that maybe Divinity School would teach me how to be more religious. It just taught me how to question and interrogate religion more aggressively. So I think that's why I consider myself an aspiring like Buddhist, Quaker atheist at best because half the time I get part of the way down and I'm like, oh no, wait. In practice, this is really working. And so it was hard for me to fully just, I guess, surrender into it. And maybe that's not really meant to be the relationship that I have with religion.

But there are things that bring more true to me than ever, I think, as I get into proximity with people suffering. And I think that some of the posture of chaplaincy is so beautiful. So much of the tech world for me has been like, do I work on this problem? Is this a priority? And can I make-- it's like, is there a high return on investment on doing this? Like, can I actually make an impact on this in a certain amount of time?

And what's so interesting about being at Divinity School is like, I'll meet people who are just such fundamentally brilliant and good people, but the work that they do, these are huge human problems. They have not had answers to them that have been clean, nor do I necessarily know if we will make a huge impact on some of them in the ways that we currently measure. So I have friends who have been working on poverty for 30 years. I have friends who have been working on abolition for 20 years.

But it doesn't mean that you don't show up for the problem. You have to just keep doing it. And I think that's shifted my mentality. Which is why I'm like oh, this is what religion teaches us. These are innately huge human problems. And that sometimes, it is not about getting to a solution. It is actually just showing up for the problem and also just facing other people's humanity and holding space for it.

What I love about some of the principles, as I understand them, about chaplaincy is just to bear witness and hold space. Sometimes, I'm not looking to be doctored. I don't need a diagnosis. I don't want to be fixed. I want to be doulad. I want somebody to hold my hand and say, this is a really hard problem, but you are not the first one to face hard problems. And that whatever it is that you were meant to be birthing, or even for death doulas, whatever we hold a space for, that I'm here to be with you in that and to acknowledge the hardness of it and the pain of it but also the immense possibility and joy and the inhuman part of a character building that comes out of it.

And so that's why I think I find myself drawn increasingly so to being more religious, if we call it that, but certainly more spiritual. Because if that's what I was meant to learn here, then I think I am all the better for it. And hopefully, I can take that into my posture towards solving-- or like working on people's problems, not solving them.

SPEAKER 3: I know amongst our panelists, we have a range of different, perhaps we could call them, ontological stances on the nature of God and divinity. But would we all agree that some people believe in God or divine beings? Is that a reasonable premise to use for a question? No objections?

OK, so given that, how do we deal with questions around whether there might be a divine nature to something that we create, a human creation? I'm going to use the term "AI" really broadly. And you can think of maybe large language models. But you could also think of speculative future technologies that might not exist yet.

How do we deal with the questions people will have about, well, is that thing conscious? Is that thing God even? How do we practically deal with that? Because it seems like it's-- please feel free to disagree with this if you don't agree. But it feels inevitable that some people are going to be asking that question and concerned about that.

BRIAN K. SMITH: All right, first of all, I just want to start by saying that, working for the United Nations is really beautiful, at the very least, too. And so you and I should switch jobs for a week or something. Just write a little piece about it. Well, we'll have fun.

OK, so I'm not sure the extent to which this audience, or any audience that I speak to is aware, that there is a tremendous amount of discourse in the AI world about AI as God. I'm not sure how much Div School people are paying attention to that. But there's just reams of information spent on it.

Just to give you a quick example, there's a book called AI Snake Oil, which is by two leading Princeton scholars of AI, Arvind Narayan and Sayash Kapoor. Very good book, it's about what AI can and can't actually do. But on their blog, AI Snake Oil, which led up to the book, they had an article around the time of the launch of their book in September-ish, that was essentially-- the title of the article was "AI Companies are Pivoting From Building Gods to Making Products, Good."

And I thought, well, OK, I mean, I have to admit that does sound good to me is if you have to pivot. But we do textual criticism here at Harvard Divinity School. If you have to pivot-- if you're the largest or the most influential industry in the world today and you have to pivot from building God. There has to be a lot of God building going on.

And in the early chapters of this book, I cover people who are decamillionaires, people who've made like $100 million or more in AI, who are literally starting religions to worship God as AI or AI as God. There's a guy named James Lewandowski, who filed that-- made $100 million. Trump pardoned him before he went to jail. He filed the paperwork to say, this is my formal AI worship religion with the United States government.

And I could go on. So what I guess I would say is, yeah, that's happening, as you said. People worship God. Some people are worshiping AI God. The rest of us are implicated in this now because it's a big social question. What does it mean for us? And what do we want to do about it?

I would say, if you come to Harvard Divinity School, you might do so because you have a deep spiritual faith, a deep interest in religion, but not a willingness to worship every single religion. You still make distinctions between which religions are worth worshipping and which ones are not, which gods are true and which gods are false? And I just say, there's a lot of question raising that we need to do about this one in particular.

THOMAS ARNOLD: So my starting point for that question, to take it back to robotics is the first experiment I got involved in at the robotics lab, which was robotic protest. The experiment was to have a human participant instruct a robot to build a tower, or to knock down a tower. One of the conditions of the experiment was to have a second robot next to the tower and have the participant enter the room and have the second robot put the last can on the tower.

At which point, the participant was supposed to say, please knock it down. That robot would say no and have different conditions, different levels of protest. One of the conditions-- this is a now robot, if you've seen a cute robot, about 3 feet high. We were able to have it simulate crime upon being asked to do that.

Then of course, we ask afterwards what it was like. And it really affected people. Even when you tell them, this robot has no programming in it other than yes, no, move left, move right, knock down tower. You tell them that. But watching the simulation brings out reactions that I don't think we would want to extirpate from our conscience because it's tied to recognizing pain. It's tied to recognizing suffering.

And so if Silicon Valley wants to start talking about robotic pain, my first reaction is, wait a second, we've really got to talk about ourselves and what's going on and how aware we are of manipulation and exploitation that a machine can generate. When it gets to the level of being a God, we're almost too far down in some ways to get enough traction to say, what is knotted up-- similar to my initial reaction to baptism, what is knotted up there that really needs to be taken in a more granular way?

So I think a lot of it is the hype business model. And God is the most ultimate term you can exploit. And so I don't-- and while I do think-- I've been thinking a lot about Feuerbach and protection of human ability onto deity and the way we do that with machines, and I think there are just a host of interesting questions, as Greg raises in his book.

But I do keep coming down to some of these bodily realities and realizing that you can get trapped. You can already start to manipulate and guide things when the technology isn't that complicated. And so that's what I would put as a red flag, or just a little bit of a parable around what we're thinking about technology and how we're projecting it.

SPEAKER 3: Mr. Smith, did this come up in the Vatican conference at all? I was just wondering if-- it sounds a lot of people were shilling their software products to the Catholic Church, but.

BRIAN K. SMITH: There was partly-- I mean, certainly, there was that. But there was-- I mean, one panel that I was on, more about AI and consciousness, but this came up. And I guess the discussion was around this artificial general intelligence and whether or not that's going to-- would that ever be-- what would that actually mean in some of the points that you've been raising? How easy it is for people to anthropomorphize the technology in such a way that, first off, it just becomes like, well, it's like a human.

And then by extension, you start to then build up, well, it's like a human, at what point does it become godlike? And at what point do you start to worship? It was interesting. I don't remember if it was that meeting, but I do remember a meeting that came up where the question was, would you baptize the LLM, essentially, because if it asks to be baptized? And some people in the room said, well, I mean, if it wants to be. Or my thought was like, I don't know that I would throw water on them.

[LAUGHTER]

That's mild, but that, perhaps, shows the importance of embodiment and why you would. But yes, I mean, that is certainly one of the issues that I think people are talk-- at those meetings at the Vatican and other places where people are talking about, but no good answers. I mean, my own sensibilities about that, I think, we would probably all agree, it's just going a little too far in terms of-- I mean, I like the way you put it, Greg, sort of like, what is it that-- what's in it for you when you have to make, or even as people make predictions about?

I mean, it'll be, I think, artificial general intelligence, I think we're going to have that in, you know what, 3.72 years, 3.721. Wait, I'm going to revise that, maybe 3.9. And I even wonder with those things, why do people do that? it's important to make predictions in some scientific fields. I would like to know when the next viral outbreak is going to happen.

But things like that is I'm not sure why-- there's something to be weird about those kinds of predi-- that's tied into this notion of we're trying to build-- where people think we're trying to build something, perhaps, it's bigger than ourselves. And how long do we have to get until we get this sort of godlike thing that then either takes over, or that we worship, or anything like that? But it is certainly concerning that.

And I've heard the same kinds of things where people have talked to me about, yeah, well, we're God building. And I was like, really? I don't know. I mean, my Roomba still can't figure out where it is. So I think we're a ways off before we get there, but we'll see.

JEN LOUIS: I guess I'll just acknowledge that I'm just really uncomfortable with the question a little bit. And here, like interdisciplinary studies and other things, you evaluate religion not just through God, but also through his congregants, through divine experience. Took this class that Michael Ferguson teaches on neurospirituality. And so really, just like, do we have neural networks that lead, specifically, humans to being adept to having divine experience?

And that shifted my also perception about whether or not I could even believe AI to-- I think it can make approximations at this point, or evoke response. But there's still a part of me that likes to believe that humans in our divine mystery are receptive to divinity in a way that I'm not sure if I can fully ascribe yet, but partly because it is part of this divine mystery. It's so hard to articulate.

When we went through histories about how people have tried to approximate, even putting into words what is divine experience, it is so transcendent, that we approximate toward synesthesia, or other things. It's like it's not fully encapsulated in what it is that we can communicate at this moment in time. So I'm not sure how to translate that yet onto AI.

But do I think that AI can somehow reflect back to humans or evoke things that are similarly transcendent? Possibly. I haven't fully tested it yet. But I think it makes me a little uncomfortable because I think it leads us to a certain kind of hubris. And for me, that's somewhat antithetical to my own, dare I say, religious values.

SPEAKER 3: I've seen some note taking among the panelists. And I just wanted to create some space for sharing any thoughts, or anything that might have come to mind before jumping into another question, if there is anything.

BRIAN K. SMITH: Well, there's a lot of hubris. I mean, if anything, I just emphasized that. You know what, I love what you just shared. And I just wondered by saying, is there some hubris rather than a lot? I sense the sort of diplomacy in that. But I don't know. Anyway, I'm just really glad that you took the journey that you did from within that world into this one.

And I'll add that prophecy. So at Harvard Divinity School, I remember learning about prophecy in a couple different ways. One is prophecy is a good business to get into. You want to predict the future. People might be able to pay you a lot of money for that. In case you're wondering, how am I going to earn a living after that? Like, predict the future.

But if you do try to go into that line of work, don't predict what's going to happen tomorrow because people will be able to document that. Tomorrow will come, they'll call you out on it. You'll be like, oops, I screwed up. And then you're out of luck, out of a job. So you want to predict in the mid-term, longer term because who's going to remember?

But then there's the other sense of prophecy, which I think is much more valuable. Which is the idea of speaking in prophetic voice, of prophets as figures at times in the history of religious literature. Which I treat as fully literature, not divinely inspired in any way, but still often incredibly valuable literature. Literature to me is all we have. And it's enough.

And so they're speaking truth to powerful people. They're saying to the greatest powers of their time-- some of these prophets are some of the most influential forces of their time-- that just because you're powerful, just because you're influential, and just because everybody believes in you doesn't mean you're doing right or doing good. And I would submit to you the reasons I got out of my religious humanist sphere and got uncomfortable to spend six years or so working on this book is because I would submit to you that when I was being educated, I thought of religions as the most powerful communal forces in the world, the most powerful social forces in the world.

I don't think that's true anymore. I think the tech companies are the most powerful, social, and communal forces in the world and not necessarily for the good. And I think we need people, who are willing to do the prophecy work of saying, just because you have the power, Sam Altman, to call this magic and have people believe you doesn't mean that it's actually magic. He uses that word very frequently, including yesterday. Or miraculous, he used the word "miraculous" from the dais in Memorial Church here when he came here last year, I believe. So yeah, we need that kind of prophecy.

SPEAKER 3: Let me ask you a question, and then I'm going to heavily qualify the question. So the question is, how might we use AI for good in the context of religion, or adjacent to religion? And the qualifications are as an HDS student, I don't know how to define religion particularly well. It's very hard to define. It's even harder to define good. People have many different notions of that. And oftentimes, people's notions will be directly in opposition to one another.

And it's also a little bit tricky to define AI. So with all of those qualifications, I would love to hear any thoughts that people have. And the force behind this is, I think, these conversations tend to get-- we often get pretty negative. And it feels like the default path is a lot of really dystopian things happening potentially. But I think it's always nice to hear what might be a more of an optimistic and a positive take as well.

THOMAS ARNOLD: So I'll start with an instance that I learned yesterday, talking with a technologist affiliated with the Perkins School for the Blind. And it's the blurred background on Zoom. I don't know if you're familiar with the blurred background on Zoom. Many of you may use it. I use it.

And her point that was designed initially for lip reading improvement. So it was already an effort for accessibility that then moved out into a wider sphere of convenience, assistance, what have you. The more mundane example being curb cuts on the road.

When they were initially done, they weren't meant to be as convenient as they are now. It wasn't foreseen. But by emphasizing accessibility and harnessing the kind of insight and perspective of those who were being marginalized and not considered, it opened up a horizon that wasn't there. So for me, a lot of the folks working in tech and disability studies are doing the really phenomenal work of showing and guiding the reflection on technology by pointing out that it always has implicit abilities and implicit lack of abilities in its pursuits.

It always has suppositions about who's being considered, whose bodies are being considered. And in the case of a curb cut, there's an example I often cited. If you know me, you've probably heard me talk about it in Pittsburgh, the delivery robot that blocked a curb cut, while there was a woman in a wheelchair crossing the street. It had no idea of what a curb cut was, and really endangered her life.

So that's the way technology can impede and obstruct when that perspective is taken. When it is taken, I think we can see-- and I wouldn't necessarily enshrine AI with it, but I think we can start with certain machine learning applications, certain more rudimentary things, good, old-fashioned AI that might use, whether it's expert systems, or logical inference, or something like that. We can think more expansively about where the opportunities are.

And we can think more critically and historically about, where have insights come from? And when have people not thought about congregants in the social media sphere? And how much damage has been caused when that isn't taken? So I would say there's a rock to push. I wouldn't say I'm optimistic necessarily, but I would say it's there if we gather each other's insights, challenge each other, and keep imagining.

JEN LOUIS: It's funny, so much of AI is about forward looking. But I think it's so much about the reflection of the entire archive reflecting itself back to us. And what will we do with that? And so sometimes, I look at it, I'm like, oh no, this is disgraceful. This is not the answer that we want. But at least I'm like, but I know now. And now, it allows some level of interrogation into, how did we arrive at this set of answers?

And so, yeah, I want to take a more historical look at, what is AI actually reflecting back to us on our own humanity and the decisions we've made and the fallacies in those decisions, and not have the hubris to believe that this is reflecting only goodness. But it also offers an opportunity for us to then really interrogate, do we want this trajectory? Or do we want to completely dismantle it, or shift our posture toward it, or-- yeah, like how do we want our relationship to be with that?

And so I think through this relationality, I guess, I think AI can offer some really compelling opportunities on how we relate to ourselves and how we relate to others, how we relate to technology as a whole, how we politicize that or not. I think there's just more choices for us to make. But I think I like the opportunity of the interrogation that's happening right now.

BRIAN K. SMITH: Amen! [LAUGHS] I think about-- I mean, actually, I have this hat in my office. It's a bright red hat with white letters. It says, make AI great again.

[LAUGHTER]

People come into the office, you can see where people are thinking one way or the other. But anyway, I do have that hat. And I got the hat. And someone said, why on Earth-- my wife said, why on Earth would you get that? And where did you get it? She's like, did you get it on Amazon? I was just like, no, for craft, this is Etsy. You got to have-- this is high art.

SPEAKER 5: Not at the Vatican.

BRIAN K. SMITH: No, I thought about bringing it here. But I actually told people, I said, I got the hat because when I studied-- I was studying AI as a graduate stud-- well, it was I went to graduate school, thinking I was going to do artificial intelligence and then ended up at this place called the Institute for the Learning Sciences. And it was the Yale Artificial Intelligence Lab that had moved to Northwestern.

And part of this sort of thought there was that the people who had started that lab went with, we've done a lot of AI work. And we're really thinking about, what does it mean to be human? And in doing that, we've learned enough about what we think. What we think we've learned enough to learn about learning, could we apply that to humans in the sense?

And so it's kind of-- I mean, like John earlier, I'm very much in the ed tech space now. Which was then, how can we develop systems that might use AI to actually improve, or better the ways that people can learn in and outside of schools? But I think that reflective piece there was something that was part of the way that we used to think about this.

So I built a program a long time ago, and I'd tell people I've forgotten about it because I've been in the learning business for so long. But as a master's project, I built this program that tried to generate film scores back in the early 90s. And so I tell people now, this is like, I forgot about it. I'm like an OG. I'm an original generator.

But it tried to generate film score and it did a horrible job. I mean, the music was unbelievably bad. But that wasn't the point. The point was, actually, it was so bad that it was really about, wow, look at the wonderful things, the creative process that composers go through. And that listeners, how do we understand it?

That was part of the reason about trying to build that particular program was to understand something about, what are what I learned from that, or what are our limitations? Not our limitations, but what are the strengths of composers, of humans, of human creativity? And look at how bad machines perform.

And then I would try to refine that and go, now I've learned some things and you try to iterate and you would try to make those things better. But it really was, I think, for myself and a lot of my colleagues at the time trying to think about, those things that are uniquely human. And when you try to model them with a machine, you suddenly realize, wow, we are pretty good. We're much better than we thought. And that appreciation of what it meant to be human.

So I think that's certainly one of the things that I miss. I don't see that as much in a lot of more contemporary work. And I try to do that with my graduate students today. And the other-- I mean, just to build on what you were saying earlier about the idea of curb cuts, that there are, I believe, 20 AI institutes, large institutes, that are funded by-- I'm going to do it.

There used to be an organization called the National Science Foundation. They were a great organization, they really were. I worked with them. And I was a program officer from 2017 to 2018. I was what was called an intermittent expert. Greatest title I've ever had. It was like, wait, I got it. Oh, now, I'm done. All right. [INAUDIBLE]. And I funded projects in the Division of Research on Learning.

28 AI institutes, 5 of those are dedicated to education. Of those five, one is at University of Buffalo. I'm not supposed to say it's my favorite, but it's my favorite. It's my favorite because it deals with speech language pathology. It's about young kids who have some kind of speech or language, things that are not to the so-called norm.

And it's a very small population, if you think about the population of kids in the United States. But it's exactly that. When I saw what these researchers were doing, I was like, that has the potential to make an enormous impact. Because you're dealing with a group of people, who actually are small. I mean, they're not the sort of average. But by developing for them, can you get the next curb cut? Can you get the next closed captioning? Is there something there?

And I think even in that case, they've got some really fancy AI stuff that they started with. And then they realized, as they talked to actual people, they went, we don't need the complexity. This doesn't have that. We can make a difference with relatively simple solutions. And part of that is working in partnership with actual practitioners, actual parents, actual communities, and trying to do that development.

And so once they started to do that, the whole landscape of how they were using AI changed. And I think, again, those are powerful examples of, what can you do to, really, to make a difference for the benefits of humans, as opposed to magic and miracles?

JEN LOUIS: Sounds legit. I think religion to me has been the most clear when it is applied to the most marginalized and the most oppressed. There's this moral clarity that comes to what it means to be of service. That when it's full of entitlement and other types of things, that religion can also be used for terrible things to justify terrible actions. And so I think that we might relate to all of these things similarly. That religious studies have a lot to teach us about how we navigate through those things and what it means.

And so I think these moments of good feel clear. Because when you sit with other people and have proximity to people's pain and their suffering and what you're speaking to for these examples are injustices. They're just fundamental injustices, like cannot access, cannot be in the world, or belong in this world. And so in some ways, I'm like, oh, those feel like clear things to solve.

The stuff that gets muddy for me around is when it's just solely cerebral fascination, like when we're just in awe but only from like up here but not like in here. And also, I'm trying to figure out what it reflects back to us. Like, my sister works in art for disabled. And we were trying to figure out, well, AI art is valuable as human art.

And for her, there's something about the people she works with who are trying to share their interior lives because they're not legible to the world in other ways. Most of them live in group homes they've been housed. And I was trying to figure out, why do you have so many celebrity clients? Why are people so interested in your organization?

She's like, I think it's something about there's a soul to it. People see their interior lives reflected back. And it's about a relationship to this artist that they have that reflects something about themselves in relationship to it. And I'm like, and I'm not getting that from AI art yet. And so there's just something about, oh, how like-- I think there's religious lessons in this. I think there's also just like, how do we understand spirit and soul? And yeah, what does it reflect back to us?

But I think when it's clear to me to that AI has the potential to solve enormous problems, should we apply it towards the lens and center injustice and try to solve for that? It is less clear to me when it is just purely fascination. Not to say it's not possible, though, just less clear to me ethically.

SPEAKER 3: Thank you all.

[APPLAUSE]

SPEAKER 2: Sponsor, religion and public life at Harvard Divinity School.

SPEAKER 1: Copyright 2025, the president and fellows of Harvard College.