Video: Humanity Meets AI Symposium: Closing

April 18, 2025

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

Harvard Divinity School.

Humanity Meets AI Symposium Closing February 28th, 2025.

ANNOUNCER: So I'm excited to announce our final talk. So Professor David Lamberth is going to close us off for the night. So Professor David Lamberth is a professor of philosophy and theology at Harvard Divinity School, specializing in pragmatism, religious experience, and the intersection of neuroscience and philosophy. He's written extensively on William James, and is currently developing a project on pragmatic approaches to religion. He also happens to be the favorite professor of many of my friends. So I look forward to hearing from him.

[APPLAUSE]

DAVID LAMBERTH: I generally have two things I really don't like to do. One is being the last speaker in a conference that I wasn't able to attend all of. And the other is following somebody who's talking about societal collapse.

[LAUGHTER]

On the other hand, after that, maybe, maybe really anything goes. I had quite a lot of thoughts in thinking about the theme of this conference. Humanity Meets AI. And I want to share some of them with you. I wound up writing them in an aim to maybe being a little tighter and a little briefer. We'll see how that works out. Brief is always relative. So let me start off with a couple of comments on the obvious, which is a point that we've no doubt all thought about.

But on which the conventions of usage have probably taken over for us. And that's the meaning of the word "intelligence." Artificial intelligence, general artificial intelligence, AI, what exactly do we mean when we're talking about this? AI has quickly become a proper noun that denotes a class of functions, something like the way. Coke became a generic word for soda of any type. When I was growing up, leaving people to ignore what actually the specifics of what they were asking for was.

ChatGPT is the new Coca-Cola, or at least OpenAI hopes it is, in that it's a stand-in for a host of things and functions, even though it itself is a quite specific product, not definitive of all that we actually should be thinking about when we're talking about AI. There's more to say there. But I want to focus on something lost behind all this branding and naming. And that's the question of what we mean by each of the terms. So, casually, tossed around as artificial intelligence.

Artificial could imply a number of things. Colloquially, it actually reads as fake, as in artificial maple flavoring or artificial color. Even though I think it's clear that when John McCarthy and others began to speak about artificial intelligence in 1956 at Dartmouth. For example, they had in mind the contrast between humans and computers. Humans and some kind of automata. That is something that was built by artifice.

The word intelligence in English is also something of an ambiguous term. We talk about people who are highly intelligent, for example, as if intelligence is something that admits easily to measure. Indeed, the development of German psychometrics in the second half of the 19th century, along with the work of French psychologist Alfred Binet, among others, yielded quantitative measures of so-called intelligence, such as the Binet-Simon and later, the Stanford-Binet. And then after that, the Wechsler IQ intelligence quotient tests, among others.

Intelligence on this positivistic quantitative model is something we can measure and rank people in relation to. Because of the formalizable and quantitative aspects of this, this approach yields well to the idea of machine or artificial intelligence. It has also, by the way, sat quite comfortably with eugenics and other racialized and cultural forms of hierarchical categorization. Now, that's a late 19th and 20th-century version of intelligence, which still has quite a bit of remaining cultural currency, quite a bit more than most of us are willing to admit.

When we say somebody's smart or that's a really smart comment, we're quantifying over the idea of this intelligence. But the concept in the word intelligence is far older, deeper, and more complex in the west than that recent domineering trajectory. Intelligence in English comes from the Latin verb "intelligere," which we usually translate to understand, which then, when contrasted with similar terms, is typically contrasted with similar terms in Latin, such as "cognoscere" to know and "sentir" to sense.

And this combination of terms clearly highlights having something like when you have intelligence, you have something like a conceptual or maybe just a higher order grasp on things, not just being able to respond to a particular question with some accurate facts or data. Literally, the Latin suggests between the words. And thus intelligere comes to be what, for example, Augustine or Anselm of Canterbury in his famous argument for the existence of God.

Intelligence, in some sense, is what both of them are seeking, as they both sure in their faith, nonetheless seek something more, something different, perhaps higher, maybe deeper. Fides quaerens intellectum. The Latin is from both of them. This is usually translated into English as faith-seeking understanding. But you can see from the Latin that we might also say faith-seeking, intellection, some dynamic action of intelligence.

Now, understanding the English gloss for intelligere comes from its German antecedent, Verstehen, which later is replaced and maybe clarified by the more common word Verstehen. Both of these emphasize standing with Verstehen emphasizing standing for. So understanding indicates a standing for something, or perhaps better, a judgment of what something is standing for.

In the late 19th century, German philosopher Wilhelm Dilthey took the notion of Verstehen, or understanding, as central to his distinctive approach to what the human science is. A combination of the humanities and social sciences were trying to do or should be doing. Dilthey took people like the positivistic IQ measurers, I was mentioning earlier, to be typical of a natural science reductivist orientation focused on quantifications and explanation.

And from these, and these people would take these explanations, and quantifications, and model the world as some closed system. Seeing something deeply limited in this approach relative to human lived experience, Dilthey instead sought to develop the role of understanding of verstehen in conjunction with the imagination, in order to show both how lived experience, lived human experience is constituted, and how we should approach its complexities.

Taking the psychic whole as the object of interpretive study, Dilthey insisted that there is purposiveness in human beings lived experience, a combination of drives, interests, and orientations that can only be approached interpretively through application of a method seeking not to explain, but instead to understand. So, here, again, we have verstehen, intelligere, understanding. But this time, overtly attentive to the nexus of complex human lives that are both inner and outer and social all at the same time.

This use of intelligence might suggest something that could maybe be modeled and perhaps imitated. But Dilthey's key takeaway is that there's something about the meaning of human lived experience that is antithetical to reduction and the explanatory approaches of the hard sciences. Jumping forward a bit more, I want to turn to the early period of the development of computers and computational systems.

In 1967, Hilary Putnam, soon to be become one of Harvard's noted philosophers and later one of my teachers. Putnam gave a paper titled Turing machines with reference to the cryptographer and British genius Alan Turing, in which Putnam proposed a view that came to be known as functionalism or Turing machine functionalism. The idea of that paper was that the brain is essentially a piece of hardware that's running a set of functions that you might think of as software, and which could, therefore, in principle, be replicated and run on some other hardware.

Though the hardware might be distinctive in the case of the human, it's the function of the software that is valuable in terms of knowledge production, and thus the mind as software is really the takeaway. This is, I should note, not at all what I was thinking about. Putnam's functionalism caught on quite quickly with cognitive scientists and also with some philosophers. And it inspired an extensive amount of work in computer science, as well as early developments in artificial intelligence at MIT first, probably, and then other places. Putnam was at MIT at the time.

The idea of functionalism is what goes on-- the idea of functionalism is what goes on in our brains and hence, in our consciousness, our lived experience, and really our lives such as we have the mentally complexity issues aside. The idea is that can all be replicated at either the individual level or at scale. So, per functionalism, if you built a machine that gave responses, it's indistinguishable from human, in all cases, the machine would thus be passing the Turing test.

And then there would be, per functionalism, no reason to think that it wasn't doing just the same thing as human beings, mentally speaking. And therefore, it should correctly be entitled or called intelligent. Such a machine then, would be potentially interchangeable with human beings. Cue the robots coming for all your jobs. AI replacing all sorts of human functions, even the AI ChatGPT intimate partner, which, by the way, the radio program, the daily podcast from the new York Times this week did a story on. It was not as disturbing as they hoped it was, but nonetheless troubling.

Philosophers criticize Putnam's functionalism. And Putnam himself came to renounce it in the 1980s based on realizing that the conditions for making actual meaning and reference, usually in his cases with regard to sentences, but generalizable from that as well. The actual conditions for making these meanings in reference can't simply be reduced to something akin to software, whether in our heads or in other hardware media.

John Searle's famous Chinese Room example, which many people know, which was developed in part as a rejoinder to Putnam's functionalism, also showed that being able to produce the right answer or output to a question didn't actually imply that the person producing it actually had any understanding of its meaning. Rather, you're just producing the right sequence of sounds and words. While Putnam didn't fully agree with Searle or many other people, they both agreed that functionalism should be abandoned.

But the idea that what it is to mean and understand can be fully rendered through software has proved quite sticky. And we default to this idea often when we're thinking about AI. Technologist and innovator Ray Kurzweil, who was mentioned earlier this afternoon. Kurzweil's vision of downloading his consciousness and upgrading his body before he dies so that he might be immortal depends in its basic articulation on a version of the functionalism thesis.

Kurzweil refined this view a bit over the years in hybrid ways. But the key commitment to functionalism to the idea that software or some alternate hardware can simply produce meaning. And more importantly, the same meaning as humans previously had with some continuity or iteratively have. This idea still at the core of many versions of techno futurism.

Now, I'm inclined, when I reflect on these directions in AI and futuristic prediction, to think that very much as tilt, I wanted to point out. We're often quite quick to oversimplify not only the human mind and our own consciousness, but also the depth of the importance of the embodied social realities that constitute who we are day-to-day.

It's convenient to think that the telos of all that we are is just intelligence in some positivistic sense. And that if we could simulate that through whatever media will have captured what's distinctive about ourselves. Putnam came around to seeing that things were more complicated than that typified in his famous anti-functionalist and anti-reductionist rejoinder that meaning ain't just in the head.

That comment related to particular views of language. And it's a little bit abstruse to the drift of my argument. But the idea that in reproducing, say our ability to crunch data or to use language in convincing ways, we're just making the human intelligence that we simultaneously value. That idea is the one I really want to underscore as something we should be questioning.

Let me turn for a moment back to the to the ChatGPT of the world. Admittedly, an arbitrary non-representative sample of no doubt what's been being discussed over the last two days. But let me turn to that to see if I can illustrate what I mean. When you engage one of these GPTs, you're working with something which has a vast access of information, far more quantity-wise than any human being could have as an immediate resource.

But the caveat with the chatbot is that you, the interrogator, ultimately have to judge the quality of the output that it's giving back to you. This is so for a couple of reasons among them whether you or someone else has sufficiently defined for a brief term, the point of view, the interests, and maybe even the values that the program should be trying to represent as it accesses and construes the data.

But it's also because the chatbot is fundamentally and structurally unable to judge the quality of its own data, unable to judge whether its data actually has the meaning that it seems to have, and certainly, the meaning that it would have were we to assert it. So it says with all matter-of-factness that I, David Lamberth, am a specialist in American pragmatism. And that I'm the author of this or that paper. But then it adds 2 or 3 I didn't write. They all sounded plausible, by the way. But I didn't actually write those. No one else did either.

All that is obviously a function, you say, of it being insufficiently constrained to a reliable data set. True enough. But what constitutes knowing what a reliable data set is and how constrained to it one should be? Perhaps, these are simply issues of complexity. And the need for iterating in the software, which would eventually take care of them. Indeed, with ChatGPT, good prompt engineering, and iteration can get rid of quite a bit of this.

But I suspect there's something else at work in the disjunction here that derives from features of human cognition, consciousness, and experience. And that has to do, in part, with the biological embodiment of our minds and our experiences, meaning, and intelligence, which is located in individuated selves, experiencing in individuated bodies, able to refer to worlds that we concretely experience.

Neurologist and cognitive scientist Antonio Damasio has written a number of books over the last 25 years that draw attention to the role that feeling or affect plays not only in our consciousness and our mental lives, but in the very making of the brain over evolutionary time. In one of his recent books, the strange order of things, Damasio offers a very brief argument at really, as an aside, as to why he doesn't think something like the Kurzweil, and singularity, or the functionalist desire to download and offload our minds to silicon or other artificial media will work.

Damasio's point is that all of our consciousness and our unconscious brain and bodily activities, and indeed all our experience, is modulated among other things, by feeling, by affect, and that the medium of having those feelings, what both produces it and what sounds through it, or it sounds through, that medium actually just is our physical bodies and requires the particular physical constitution that our bodies are made up of.

Affect on Damasio's read indicates biological valence. And this valence itself is tied in with our attempts not only to regulate our own lives, to maintain homeostasis, to stay alive, but also to orient ourselves towards increasing the positive value for our own individual biological organisms, and sometimes by extension, to our wider groups. Damasio sees this going up to the size of tribe, and then having a hard time reaching beyond that.

Now, Damasio's view is complicated. He's bringing together all this neurology, and brain science, and also together a theory of the self, and how consciousness works, host of other things. Great fun to read. But the takeaways are fairly easy without going into all that to convey. First, our minds and bodies have an idiosyncratic point of view, each of them. One which is determined by the outer limits of our actual bodies and modulated through our bodies complex neurological and chemical systems, not just our neuronal brains or our electric conveying brains.

So it's not just all in the brain, even if it's chemically in the brain. It's actually the brain and body combined, which is required to produce the valence that we're depending on. Second, the materiality and sociality of that body with its brain produces not only thoughts and actions, but also value in the form of valence that is positive or negative feeling, which is related to each body's discrete, distinctive reality.

We're all humans. But we're all also individuated, and individual selves, and individual bodies. And thus our valence is very. Our points of view differ. And our judgments and actions follow individual suits, as well as general social trajectories. The condition for all this meaning-making is necessarily mediated through the biologically complex organism that we have evolved to become, and which has evolved into consciousness.

And as a result, Damasio thinks if you want to simulate our consciousness, you'd really need to reconstruct down to the cellular level, the actual biological body that somebody has. So we might be able to build a place for ray Kurzweil to go. But we probably wouldn't be able to get him there. Because the one body isn't the same as the other, even if it's structurally identical.

The takeaway here isn't that such simulations couldn't be done, or that AI systems can't replicate, and exceed us in many of the things that we do. Obviously, they can and already have. And to the extent that we think well about what those different systems, different strengths are, and what their limitations are, we can adapt and utilize them to do potentially valuable, even potentially amazing things.

But there's something about many of these systems as we built and adapted them so far, it's just slightly off. Whether it's the mildly disturbing superficiality of a chatbot friend or intimate partner, whether it's the heartless quality of writing that you just have a vague, but pretty sure sense of when you read an LLM-generated paper handed in by a student, or whether it's the inability to be able to know what is meaningful to focus on beyond the statistical probability that is so in the wheelhouse of our current machine learning and quantitative systems, that is, they can show me the probabilities and the instantiations, but they can't tell me which ones of them actually really have meaning relative to the question I'm asking.

Now, admittedly, even though we've made tons of progress, particularly in the last few years, we're still at the very beginning of this whole set of innovations that we're calling AI. Even at this early stage, it's clear how remarkable and potentially world-changing and valuable they might in fact be. But I want to suggest that we should keep in view, however, that these judgments that I just made about where we are, and how valuable they might be, and what they might be good for, are all synthetic and, crucially, only judgments that can be made relative to a particularly assumed point of view, one that I'm making relative to the meaning that I am individually, and that we collectively are making and feeling in relation to these actions.

These judgments are also all only available by virtue of the peculiar mix of my own experiences. Insofar as they're my judgments. They depend on what I've read and what I've thought about, who has said what to me, and when, and how I took that at the time, and what kind of feelings I had, what affect went along with them, which is part of what makes them memorable. And how the pulse of my own selfhood was, as Damasio or William James might put it, was shaping that valence and the iterative judgment thereof of value.

Immediate luminousness, philosophical reasonableness, and moral helpfulness. Those William James wrote, are the only criteria of judgment we ultimately have to lean on when we're judging value. He calls it making spiritual judgments. And then references the German, which also shows up in delta. For now, at least, it's apparent to me that the AI-- sorry, it's important, from my point of view, to suggest that the AI optimists need to look past the novelty and possibility in it all towards some additional different human questions, questions fundamentally of value for both groups of individuals, but individual, but also single individuals.

And it also seems to me that the pessimists need to look beyond the failures and inadequacies to the reality of what achievements and changes artificial intelligence and broader sense is already making and will make. Crucially, as human beings with valence points of view and individual and collective values that are deep products of each of our own lives, we all should think seriously about understanding, about really understanding in the classical sense of intelligence, all of what's going on. And thus think about what AI means for advancing our human values and meaning, advancing our humanity.

The technology is difficult. But the value proposition is really the important challenge. Thank you.

[APPLAUSE]

Sponsor, Religion and Public Life at Harvard Divinity School.