Is "Knowing" More Than Neural Networks? Christianity, Scientism, & AI | Dr. Ian Hutchinson
1 hr 3 minAudio
How do we really know anything - and can an AI "know" at all? In the third annual Dallas A. Willard Lecture, MIT physicist Ian Hutchinson takes on scientism, the belief that natural science is the only route to real knowledge, and asks what it means for Christian faith in an age of artificial intelligence. Drawing on three kinds of knowing (knowing that, knowing how, and knowing someone), Hutchinson argues that scientism is an intellectual mistake, that the "science vs. religion" conflict story is a myth, and that large language models simulate thought rather than possess it. His conclusion: the image of God lies not in raw intelligence but in our calling to know and love God, and love is more than neural networks. Following the lecture, philosopher and epistemologist Tony Bolos, Director of New College Madison, offers a response on justification, Plato's Meno and Euthyphro, and why we can know things (like love) that we cannot fully define. Ian Hutchinson is Professor Emeritus of Nuclear Science and Engineering at MIT, an internationally recognized plasma and fusion physicist, and author of "Can a Scientist Believe in Miracles?" He has spoken widely on science and Christianity with the Veritas Forum, InterVarsity, BioLogos, and the American Scientific Affiliation. In this lecture: Three kinds of knowing: propositions, skills, and acquaintance What makes natural science reliable, and where its methods stop Scientism and positivism, from Comte to Russell, Monod, and Dawkins Why the "conflict" between science and Christianity is a historical myth Reading Genesis and the "two books" of Scripture and nature Can AI know facts? Can it know how to do things? How large language models work, and why agentic AI calls for regulation Artificial intelligence and the image of God Love as the highest form of knowing Recorded October 16, 2026, in Madison, Wisconsin, as part of the Wisconsin Science Festival. The Dallas A. Willard Lecture is presented by the Lumen Center, a ministry of the Stephen & Laurel Brown Foundation, and is generously supported by Fieldstead and Co. Learn more: https://slbf.org Event page: https://subspla.sh/jbfyxxd Send us Fan Mail Upper House Commons gathers the university community for spiritual, intellectual, and vocational formation. We explore big ideas and engage in conversations that matter within arts and humanities, justice and society, leadership and vocation, science and technology, spiritual formation, and theology. Whether you are a student or faculty member at UW-Madison or beyond, working in the marketplace, or serving in the church, we see you as part of our university community. Gather with us for one of our programs - our “commons” - each a pasture for shared spiritual, intellectual, and vocational formation. Head over to our events page to see what's coming soon, or mark your calendar for these upcoming programs. Find out more slbf.org/upperhousecommons
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
SPEAKER_00And thanks to all of you for being here. Um, I'm remarkably much more relaxed tonight than I was about a month ago when I was standing on this stage. For those of you who were here for my uh Can Science Be Trusted talk. But tonight I have the pleasure of introducing our Dallas Willard speaker, Dr. Ian Hutchinson. Yesterday marked the opening of the 16th annual Wisconsin Science Festival, which this year runs for 10 days with events all over the UW campus.
SPEAKER_00The theme this year is us exploring the science of who we are, of human biology, and of our place in the cosmos. So we're particularly delighted then that the Lumen Center can contribute to Science Week by bringing in a scientist of the stature of Ian Hutchinson. Ian is professor of nuclear science and engineering emeritus at MIT, the Massachusetts Institute of Technology. He is an international expert on the physics of plasma, and I don't mean the stuff running around in your veins, but that high-temperature ionized fourth state of matter.
SPEAKER_00And he's an expert on how to confine it to derive energy from fusion reactions, the very reaction that makes stars stars. Ian directed a national fusion research facility, authored 250 scientific journal articles, and wrote two advanced science textbooks. In addition to his academic work, Ian has also written and spoken widely on the relationship between science and Christianity. He's appeared in many, many different events sponsored by Christian organizations such as Intervarsity Christian Fellowship, the Veritas Forum, the American Scientific Affiliation, and Biologos.
SPEAKER_00He's also been featured in the wider secular media, including the American Association for the Advancement of Science, broadcast TV debates, syndicated radio interviews, public lectures, and international news blogs. His book, which is available over there by the reception desk, Can a Scientist Believe in Miracles? addresses questions that he's been asked over several decades about explaining science and Christianity to the university community.
SPEAKER_00Ian and his wife Fran have two children and four grandsons. He is a keen classical singer, squash player, and most importantly, Flyfisher. So join me in giving a warm welcome to Dr. Ian Hutchinson.
SPEAKER_02Thank you very much. It's an honor to be here, and uh particularly an honor for me to be giving a Willard lecture since he's someone that I very much look at look up to. Is knowing more than neural networks? Christianity, scientism, and artificial intelligence is my title. So knowing, what what does it take to know? The word has at least three different types of meaning.
SPEAKER_02Knowing that is about facts, I know one plus one equals two. Knowing how to is about skills and disciplines. I know how to read. Knowing someone is about relationships and acquaintances. I know my wife. And Christianity involves all three types. We say that we know things about God, that he is a Trinity.
SPEAKER_02We know things about Christ, that he died for our sins and rose again on the third day, and about ourselves, that we are sinners alienated from God until redeemed by the blood of Christ. These examples are propositions. So knowing that is knowing propositions. We also see in Christianity the importance of knowing how to do things, how to pray, meditate, how to read, praise, testify, sing, forgive. These are skills and disciplines that undergird the Christian life and are developed over time and by practice.
SPEAKER_02But the greatest commandments, according to Jesus, are about relationships. Love God and love your neighbor. This is a call to know and care for someone as a matter of direct acquaintanceship. And acquaintanceshi knowledge also can extend to non-personal things, for example, knowing places or knowing directly. So my plan in this talk is to organize the lecture into three parts corresponding to these three different types of knowing.
SPEAKER_02The hosts of the lecture have asked me explicitly to say something about artificial intelligence, which is today obviously a very hot topic pertaining to knowledge. And I'll have a few things to say about the ways that AI might be said to know in these different ways, or not. So let's talk about propositional knowledge. In the academy, all three types of knowing are significant, but there is a justifiable emphasis on knowing that propositions.
SPEAKER_02And understanding how we establish true factual knowledge is called epistemology. And we have an epistemologist in the house, and you're going to hear from him later. Well, I am a physicist, and physics is often regarded as the most fundamental type of natural science. Science arrives at enormously persuasive knowledge that undergirds engineering and medicine.
SPEAKER_02Its influence on society and on our lives is immense, and it's a high reputation, which is deserved, uh, in terms of commanding public assent. Maybe it's a bit rocky at the moment in a popular perception, but generally speaking, it still has a high reputation. And that reputation and that success depends upon a specific epistemological discipline that we should pay attention to. And science, and I mean natural science, is the study and knowledge of nature, understood as the normal, regular behavior of the world.
SPEAKER_02That particular focus blossomed in the seventeenth century scientific revolution. It relies upon reproducible experiments and observations, together with a requirement of unambiguous clarity. That clarity is necessary in order to know whether or not results are reproducible. And it often comes from, for example, quantitative measurement, but it also involves other specialized practices and ways of reporting.
SPEAKER_02Robert Boyle, who was one of the founders of the Royal Society in London, wrote a whole book about the vulgarly received notion of nature, in which he listed eight different meanings of the word nature, and he insisted that for science the meaning that was important and should be used is the normal, regular, reproducible behavior of the world.
SPEAKER_02Setting aside those other sometimes confusing and very common meanings, and natural philosophy, as it was then called, as science was then called, was and is what we find out from reproducible experiments and observations and unambiguous clarity. Of course, though, many academic disciplines don't possess this reproducibility and clarity. You can't do experiments in them. So natural science is thereby limited in scope.
SPEAKER_02Other disciplines have other ways of discovering factual knowledge. Well, the methods of science proved so remarkably successful at establishing reliable knowledge that it captured the imaginations of many scholars across a whole range of non-scientific disciplines. So gradually, a habit of mind grew up which is called scientism, and that is the view that science is well all the real knowledge there is.
SPEAKER_02Scientism is only very rarely directly affirmed because, well, non scientific topics don't really use the methods of natural science, but one can give explicit examples of its claims. And here are a few. Bertrand Russell, the philosopher, wrote in his book Whatever knowledge is attainable must be attained by scientific methods, and what science cannot discover, mankind cannot know.
SPEAKER_02Or the cornerstone of the scientific method is the systematic denial that true knowledge can be got by interpreting phenomena in terms of final causes, that is to say, of purpose. That was from Jacques Monot, a Nobel Prize winner in biology. Science proves, or excuse me, science provides all the significant truths about reality, and knowing such truths is what real understanding is about.
SPEAKER_02Once you adopt science, you'll be able to cease treating the humanities at as knowledge. This was Alex Rosenberg, not me. Or finally, I mean it most simply of all, from Richard Dawkins, the existence of God is a scientific hypothesis like any other. So those are examples of explicit scientism, but scientism is in the is in the air we breathe these days. Scientism is also promoted by a confusing linguistic history.
SPEAKER_02Prior to the scientific revolution, the Latin word scientia, from which the English word science comes, was used to describe any rigorously established knowledge, regardless of whether it was empirical or purely mental or logical. And when that was the word's meaning, it made sense to regard theology as the queen of the senti. Reference to that archaic meaning is found today in the names of various academic disciplines that are obviously not susceptible to the methods of natural science, for example, political science or social science.
SPEAKER_02Such non-scientific disciplines do have knowledge, but it's not obtained by anything approximating the methods of natural science. It generally has weak reproducibility and substantial ambiguity because it's not about nature, it's about persons and societies and politics and so forth that are not susceptible to really reproducible experiments.
SPEAKER_02Also, though, recently named disciplines love to call themselves science because it gives them an aura of authority. It's a promotional strategy. The established natural sciences like physics and chemistry and biology and geology and cosmology and physiology and astronomy and so on usually don't feel the need to put science into their titles in order for self-promotion.
SPEAKER_02So let's think about some non-scientific disciplines. A good example is history. Human history is a good example of a topic that is outside the methods of natural science. How do we know that Julius Caesar was assassinated by fellow Roman senators on the fifteenth of march forty four BC? Not by reproducibility, not by experiment or any real evidence like that of natural science.
SPEAKER_02And yet, history is real knowledge. It's just not scientific knowledge. And there are many other topics and disciplines which don't possess reproducibility or unambiguous clarity, language, literature, philosophy, jurisprudence, sociology, economics, even though it uses mathematics, politics. They have ways to find knowledge, but they're not sciences in the normal common sense in which the natural sciences are.
SPEAKER_02Well, my colleagues sometimes bristle when I say this, okay, my humanities and other non-scientific colleagues. But they shouldn't, because they bristle because they're assuming that I'm belittling their discipline by the non-scientific label. Well, I'm but I'm not, because I don't subscribe to scientism. I don't think that all the real knowledge there is is discovered by science. And yet their disciplines are all too often distorted or at least influenced by scientism.
SPEAKER_02Now, um I'm going to trespass on my philosophical colleagues for a little ground for a little while, and just mention that in philosophy, uh positivism is the word often used to describe different forms of what I call scientism. The word positivism unqualified derives from a theory and religious movement founded by Auguste Comte in the first half of the nineteenth century, and it imagined that every discipline needed to make a transition from primitive non-scientific states to become a positive science.
SPEAKER_02So he summarized this in the so-called law of three stages. Each branch of knowledge passes successively through three different theoretical conditions, the theological or fictitious, the metaphysical or abstract, and finally to the scientific or positive. And he aimed to bring about that transition for the discipline of sociology. Interestingly, Comte's positivism became a literal religion.
SPEAKER_02It was called the religion of humanity. It had hierarchical priesthood, it had public services of commemoration, daily prayers, nine personal sacraments, and so on, eighty-one annual festivals. There were several notable figures in literature who joined this church for a while, and the religion was transported as far away, at least as Brazil, as evidenced by this photograph of a positivist temple still standing in Brazil, before fairly soon dying out.
SPEAKER_02In the twentieth century, logical positivism was an epistemological view that regarded real knowledge, and indeed real meaning, as limited to what could be established by either an empirical test or by being logically self-evident. Both of these positivist epistemological fashions today are considered pretty much defunct in current philosophy.
SPEAKER_02And that's to me not surprising, since there are many academic fields that I've just listed and that possess knowledge but don't obtain it by natural scientific methods, empirical methods, and the assertions of logical positivism, that you know, this this verifiability criterion of meaning fall under the under the problem that those assertions, the assertion of this verifiability of meaning, um is discovered neither to be analytic nor empirical.
SPEAKER_02So it is, by its own lights, not even meaningful. So scientism, the belief that science is the only supreme or the only or the supreme route to knowledge is a ghastly intellectual mistake. But it still persists in popular culture and especially among anti-theists who argue against religion. What's more, I believe that scientism lies behind many of the arguments in support of the myth that science and religion are and always have been in conflict.
SPEAKER_02This myth is not new, but it's also not terribly old. It comes really from the nineteenth century. Scientism and Christianity are indeed incompatible, in part because, as I mentioned, you know, scientism is a kind of religion in a certain sense. And what's more, the Christian faith has always insisted on non-scientific knowledge, witnesses, testimony, history, irreproducible events as and so on, as being part of its scope.
SPEAKER_02If science really were all the knowledge there is, then those objective grounds for belief in Christianity would be ruled out. So it is in large measure the background of scientism which we sort of breathe throughout our Western society today, that was being expressed when uh the nine-year-old son of one of my scientist friends of today from church said to his father, I can't be a Christian, dad, because I want to be a scientist.
SPEAKER_02Where did he get that idea from, given that his father was a dedicated Christian and scientist? Um A responsible adult should recognize that the conflict myth has been thoroughly debunked by historians of science. Science's founders, um the scientific revolution's founders, were almost all serious committed Christians, here are five of them, who took major efforts to understand their faith, and they drew scientific encouragement and inspiration from it.
SPEAKER_02And what's more, succeeding generations of natural philosophers likewise were predominantly Christians. One way to look at this is to think about MIT's architecture. MIT has pseudo-Greek architecture, here are some photographs of it, um, with fluted columns. It was it was built in the 1920s. But above the fluted columns, instead of having the back and alien friezes of the drunken orgies of the gods, what we have are the names of the scientific heroes of history that were considered to be at that time in the 1920s.
SPEAKER_02And some years ago, a few of us on the faculty at MIT estimated the fraction of those names who were Christians, and we came up with a it's a rough estimate, came up with 50 to 60 percent. So it's remarkable, therefore, that many people believe the myth that science and religion are and always have been at war. Scientism also confounds the understanding of the Bible, for which some strands of Christian thought also bear some responsibility.
SPEAKER_02The Bible contains an enormous variety of literature history, poetry, prophecy, theology, biography, songs, ritual, law, letters, proverbs, philosophy, and on and on. But nothing like today's scientific descriptions of the world. No one in Bible times thought in terms of today's laws of nature, which which which come from the investigation of reproducibility.
SPEAKER_02The Bible is not written using precise, objective, unambiguous expression. That's the clarity that I referred to earlier. And what's more, it contains no data or analysis that is like today's natural science. It does marvel at the world, the natural world, and speak of the faithfulness of God in it, which perhaps points to reproducibility and was one of the inspirations of those founders of the scientific revolution.
SPEAKER_02Um but it contains essentially no science as we understand it today. Concordist Bible interpretations are motivated. Motivated, I think, most of all by scientism, by a sense that somehow the Bible needs to be scientific in the way that it makes its expression, or at least scientifically reliable. But the first few chapters of Genesis, for example, weren't written to convey scientifically how God created the world.
SPEAKER_02They were written in order to convey that he created it, and that is, he is the God of all creation, not just a tribal deity like those who believe who were believed in by most of the societies surrounding the ancient Hebrews. And so to seek scientific um descriptions in Genesis seems to me to be just a mistake. Actually, um that's been known for centuries.
SPEAKER_02And the distinction between scientific knowledge and factual knowledge is part of the historic practice of the church in the form of particularly of what's called the two books metaphor. So from the time of Augustine and Chrysostom, the church has regarded the revelation of God as coming metaphorically speaking in two books, the Book of God's Words, the Bible, and the Book of God's works, nature. And both of those books require and deserve interpretation on their own terms.
SPEAKER_02And so this quote from Francis Bacon, who was one of the f folks I showed you earlier, is he said, Let no one think or maintain that a person can search too far or be too well studied in the book of God's Word or the Book of God's works. The Bible shouldn't be approached as if it were a scientific textbook. But also, and this is where I begin to m to move on, we shouldn't regard God's self-revelation to be pr predominantly knowledge of propositions, knowledge that.
SPEAKER_02Scientific explanations, even if they exist, don't rule out non-scientific explanations. They can be true at the same time. John Polkinghorn's famous illustration is the question about this boiling kettle. Why is the water in this kettle boiling? And a valid scientific answer might be because its temperature has caused its vapor pressure to reach atmospheric pressure.
SPEAKER_02But an equally valid non-scientific answer might be because I want a cup of tea. And and those are both true at the same time, okay? So non-scientific and scientific answers can be simultaneously true. Okay, now before I move on, though, to to how-to knowledge, and then to eventually to acquaintanceship knowledge, I want to make an important remark that bears on an understanding of artificial intelligence.
SPEAKER_02And that is that almost all our factual knowledge is derived from, or at least guided by, others. In other words, reports from others provide almost all the factual knowledge we know. Well, for the knowledge of people, there's a small proportion of our factual knowledge that we pick up by acquaintances and by practice, by so it comes our factual knowledge can come from the other two types of knowledge. But um essentially all the factual knowledge that AI has comes from external digitized sources.
SPEAKER_02And that knowledge, like ours, is only as reliable as its authors. But knowing the reliability of written or oral reports is um, which is essential to our assessment of whether we should believe their reports, also depends on still other reports by other people. So uh uh and what's more, it also depends on our interpretation of their reports. So in summary, the significance of knowledge, even knowledge of facts, depends on interpretations and assessments.
SPEAKER_02And the human knowledge of facts, um whether of science or of other disciplines, is a tapestry of fragile threads that are woven together so that they make a strong mutually supporting fabric. Well now we don't know precisely how AIs come up with their responses. How does Jat chat GPT work? No one actually really quite knows how it comes up with its answers.
SPEAKER_02Um but it's certainly the case that large language models depend for their assessments almost entirely on the frequency of word patterns in the written reports which they are on which they are trained. And that leaves out some of the ways that thoughtful humans assess reliability. So, for example, we often assess whether we're going to accept or believe a person's report based on a background understanding of their motivation and self-interest.
SPEAKER_02Um, and that has to do with acquaintances and how-to knowledge. So these types of knowledge are not completely sundered from one one another, and almost all of our knowledge comes from other people. Okay, so now uh we're moving on to knowing how-to. Does AI know techniques or how-to knowledge? Actually, um simple non-AI web search engines like the original Google and and others that came before it, they've been around since the 1990s, and they've been able to find influential web documents that reference a search prompt.
SPEAKER_02And in that limited sense, they know and can report a lot of facts. But what about their ability to do things? To do they know how to how to do things and have they got techniques? It might surprise you that computers have known some techniques for much, much longer. Arithmetic was something that mechanical computers were designed to do in the early 19th century, for example, by Charles Babbage. He never actually built one, but he designed one.
SPEAKER_02Um and all computers depend on logical operations, and they can do, therefore, do arithmetic with numbers. And they can be programmed to do other tasks such as sorting numbers and increasingly more complex tasks. So that's always basically been part of computing. Programming of even more sophisticated and more specialized techniques took off shortly after the invention of electronic computers. For example, the first chess programs were developed around 1960.
SPEAKER_02They weren't terribly good players. This photo of me um winning the under 18 County Chess Championship is from is from 1968. Computer programs fairly soon surpassed my skill, and uh the world's human chess champion, Kasparov, was beaten by Deep Blue in 1997. So computers could play things with well-defined uh rules very, very well, quite early on.
SPEAKER_02They were also able to do sophisticated mathematics like algebra and calculus from early on, and uh these are the dates of certain algebra systems. Um so yes, um we know even now with appropriate hardware as well as software, computers can see, hear, speak, and so on. So, yes, computers can be programmed to execute sophisticated algorithms incredibly quickly, and it's not completely unreasonable to regard that as knowledge uh how to.
SPEAKER_02AI was also, by the way, instrumental in the 2024 Nobel Prize for Chemistry for calculating the folding structure of very large protein molecules. And the the reason why they were so helpful was that they could try all the different possible combinations of the ways that they fold very, very quickly, which was at a speed that far exceeded um human. Um this ability uh um seems to me really unthreatening as far as spiritual understanding of reality, though it might lead us, uh humans, to laziness, letting the computers do all of the hard calculations.
SPEAKER_02Human how-to knowledge involves training and discipline. And AI also has to be trained, but it's less clear that uh AI can be disciplined. We'll find out in the near future. That's one of the worries about it. But there's also, I think, something rather different about human training and performance, and that is its simultaneous involvement of higher perception.
SPEAKER_02So um consider this musical event, and I'm gonna ask you, um if I can launch it, I'm gonna ask you to think about what is it? Well, it's two brass uh plates being smashed together and all sorts of vibrations taking place, it's someone bringing down a hammer on an anvil. But no, of course, that's not actually what it is.
SPEAKER_02It's the famous hammer blow from um Marlis Sixth Symphony. Um music's meanings are are not really present in the scientific description by itself in solution. They they are elicited only by the act of hearing, um, by the connection of the music to the human, and they can't be expressed unambiguously. Um and the performers are simultaneously entering into the hearing experience.
SPEAKER_02So they're not just carrying out precise mechanical actions. And I and AI really cannot have the same acquaintance participation as humans in the performance of how-to knowledge. And that seems to me, at least tentatively, as a significant difference. Okay, um so now I'm going to go into a little deeper dive about how AI works. Um before I do so, um let's think about libraries.
SPEAKER_02We never say that a library knows. I wonder why that is. I suggest it's because a library is not an independent agent that is conscious and responsible, unlike, for example, higher animals and humans, um, who are conscious and to some degree responsible. So we should ask, are AI systems independent, conscious, responsible agents? They really aren't now, but could they ever be?
SPEAKER_02Well, agents in some sense, yes, and uh that I'm going to come to that later. But independent, conscious, responsible, it's not so obvious to me. But let me first of all make sure that I'm not um seen as being um too much of uh a posturing amateur here. I will say that I am not an AI expert. I'm not involved in the development of artificial intelligence or an extensive use of it.
SPEAKER_02I mean, I I use Google, Gemini, and things like that, as you probably all do. Um my only relevant qualifications are extensive mathematical and computational and scientific experience, which enable me to understand and outline how AI works. So I'm offering mostly simple, modest ideas and questions. But I do note that even those who are developers, experts of current AI, don't exactly know how it works. So I'm not entirely on my own in this, and they cannot predict its behavior.
SPEAKER_02So first, I've said that AIs are not independent. They depend absolutely on pre-existing factors. The internet, providing electronic and photonic connectivity linking the world, the World Wide Web protocols for publishing and accessing information, the computerization of large fractions of human activity and of human literature, um, access to enormous parallel computing systems and to digitized information. All of this infrastructure is provided and maintained by humans.
SPEAKER_02Um, so they aren't independent. Well, what we think of today when assessing AI is technically called large language models. And I'm going to offer a little video explaining how they work and explaining it uh with a smile. And I'm it's going to be a little tricky to see if I can make this work, but we we have made it work, and so uh listen to this.
SPEAKER_01So you give it an enormous amount of information, yes, and then it learns how to guess. Guess? Yes, guess. Well, guess what? What comes next? What does come next? Words, uh, phrases, uh, images, even computer code. Okay, so it understands all of that? No. It doesn't understand, no, but it can explain that. You can call it explaining, Chris, but it's just a very great guess. So it's not understanding any of that. I think you're not understanding, Chris. Okay, all right. Is AI concerning? It doesn't understand, Chris. Right, but how does it not understand and still give the answers that it does?
SPEAKER_01It learns, Chris. So when I ask AI what's the capital of France, it predicts an answer. Paris. Probably. It doesn't know Paris. It's a model, Chris. It's not eating a baguette. But what sense does it know Paris in? In the sense that after the words capital of France is the word Paris. Right. And how does it learn? It teaches itself, Chris. Oh, and another AI teaches it too. Right. Does that other AI understand? No, no, Chris. They don't understand. They guess. Right. So it's trial and error.
SPEAKER_01You see, what when it gets something wrong, it learns from that. And it adjusts some mathematical knobs deep within it. So next time its guess will be better. Okay, and how many knobs are there? Millions. No. Billions. No, wait. We don't know. Well, how do you not know? You adjust them, don't you? No, the AI does that. But you said it guesses. Well, I guess it guesses what knob to adjust, so its guesses become less guessy. So you don't fully know how it works. This sounds insane. Uh how are you controlling this? Well, if I had to guess. Now you're guessing. Well, Chris, basically what we're doing is we're giving it more and more and more computing power.
SPEAKER_01What? Do you have room for a data center? No. Why are you doing that? Because we're kind of in a race. To where? Well, to the end. The end of what? Well, the race. What race? Hold on, hold on. Good catch. No, no. The race to be the most superpowerful AI. Who's in the race? Well, the world's superpowers. What why? Chris, Chris, Chris. This is a superpower race. And we're superpowering. Right. It would make sense that the superpowers are in a race about superpower. And if you if if the superpower wins the superpower race, then they will be the superpower.
SPEAKER_01Right. So the superpowers of the world are in a superpowerful race, feeding as much superpower into a superpowerful technology, sucking the world's power, all in order to win a race that might end our race. Yeah. And just to clarify, you don't fully know how it works, what it will do, and how it'll do it? Well, we could guess. Right. So the general consensus is that AI will be extremely powerful? Yes. Right? And the general consensus is that AI can be dangerous. Oh, yes. Right. And the general consensus is it's full steam ahead because we want to win that race. Yes.
SPEAKER_01Now, Chris, what's the capital of Uzbekistan? I don't know, and I'm not going to guess. You would make a horrible AI, Chris.
SPEAKER_02It's funny, but it's also true. So what are large language models? Large language models consist primarily primarily of giant neural networks whose connection weights have been adjusted by training. And this is a very small part, imagined part of a neural network where the various neurons are connected through these weighted connections.
SPEAKER_02And they've been trained with trillions of words of pre-existing text scraped from the World Wide Web to guess the next word basically in a stream of words. They're astonishing. But in effect, they are simulating speech and thought. Scientists like me actually are very familiar with the idea of digital simulation of the behavior of the world. Of course, we do it through solving equations mostly, sometimes with random uncertainties included.
SPEAKER_02But large language models are a way of simulating language, and they do it with random uncertainties based on extensive training and prior language that expresses the thoughts of humans. So they are simulating thought. What's the difference though between knowing how to speak or think and merely pretending or simulating knowledge? I think part of the answer lies in knowledge of, specifically acquaintanceship and agency, together with awareness of context and consequences, which undergirds responsibility.
SPEAKER_02Well now, the word agency is really a very topical word in AI research. It raises in effect the hottest topic in AR currently in AI, which is called agentic AI or agentic AI. And that uses, but it goes beyond large language models. First, here's a loose definition. Agentic AI is a software system designed to interact with data and tools in a way that requires minimal human intervention with an emphasis on goal-oriented behavior.
SPEAKER_02Agentic AI can accomplish tasks by creating a list of steps and performing them autonomously. So it combines large language models, which give it the uh ability of guessing language, not understanding, guessing language, uh, but it combines them with uh tools, algorithms, data gathering, and a specific aim. It's iterative and it's adaptive. So it tries things over and over again and it can change its tactics and do so in response to the results that it gets.
SPEAKER_02So it's autonomous in the sense that it doesn't report back to the human um uh operator until it reaches some kind of stop condition. So they just let it loose and it goes for however long it's gonna go until it until it satisfies something that that has been pre-dis determined by the operator. It can consist of multi-agent systems. So there's an AI supervisor that has lots of sub-AIs that it controls or programs itself, um, and uh uh and that gives it tremendous strength.
SPEAKER_02Um but its actions can never be precisely predicted. The fact that it can't be predicted can be thought of as being a strength because it allows the system to have new ideas or explore new approaches and therefore simulate creativity. But it also is a serious danger because uh of the possibility, which we've already observed, that systems will go rogue and start doing things that its instructors, its human instructors, don't want it to. Um there are good reasons, therefore, I believe, to um uh have tight regulation of AI because it has the power to cause immense disruption to human civilization, and we cannot predict its behavior.
SPEAKER_02Also, its developers have great financial incentives to push ahead regardless of risks. I mean, some of them have talked about um those risks openly, but some of them haven't. But either but even if they do, um the financial incentives to push ahead are overwhelming for them. Well, um I need to focus now on the main topic that motivates my interest in this whole question. How does AI challenge Christianity or influence the Christian faith?
SPEAKER_02I think the main Impact is is likely to be about the uniqueness and intrinsic value of human intelligence. If machines can think faster than, know more than, and behave like people, should we regard humans as having some special value to us or to God? And if so, why? If intelligence can be artificially designed, what makes man special?
SPEAKER_02Can we believe that God said, as in Genesis 126, let us make man in our own image after our own like likeness? But also what does that mean? And that brings me uh to talk about acquaintanceship knowledge. The idea that uh humans are possessed of the image of God is obviously critical for human um human uh status in the world and and is obviously also a critical part of Christian theology.
SPEAKER_02But what exactly it means has been debated for m millennia. Various ideas have been put forward, interpreting it as based on thought or language or power or skill. But if AI is smarter than their human creators, what uniqueness remains to the province of of humans? So um I wanted to get an up-to-date take on this. I mean, there's as I say, we this Christian literature going back centuries. Um, but here's what I came up with, which seemed to me a plausible um def uh definition or explanation of what's God's image.
SPEAKER_02God's image is seen in our more moral and spiritual nature, in our shared love within community, in our dominion on God's behalf, and especially in our calling to communion with God in knowledge and righteousness. We are creatures designed to know and be known by God and to love and be loved by our Maker, so said Richard Phillips in this article which was published by the Gospel Coalition. I think that's that's pretty serviceable definition.
SPEAKER_02Um and it's a little different from many of the other approaches which have focused on our abilities and our uh and our how-to knowledge, um, because what it's focuses on is our acquaintanceship knowledge. Um and and it's um uh it sets aside intelligence as being per se where the image of God lies. Instead, it emphasizes knowing and loving God. We are creatures designed to know and be loved by God and to love and bring um Him back.
SPEAKER_02So I think uh we already knew from what Jesus had told us that love is the highest form of acquaintanceship knowledge, that we should love the Lord our God with all of our heart and mould and soul and mind and strength, and our neighbors as ourselves. Umility is an essential part of that. Christ humbled himself, and that was part of his love for us. But also love is patient and kind, love is not jealous or boastful, it's not arrogant or rude, or um as uh Paul says in the towards the end of that chapter, love never ends.
SPEAKER_02So the vocation of God's image survives all depravity and damage. The poorest criminal, the most handicapped person remain in God's love. Um, love is the defining character of God and of his image in us. His love for us outlasts and supersedes our disobedience. God shows his love for us, in that while we were yet sinners, Christ died for us. His call to love him in return is a cause, is a call to abundant and fulfilling eternal life.
SPEAKER_02I came that they might have life and have it abundantly. And and none of that call depends on AI or can be erased by its success or its failure or its disaster. Um you might still be exercised with a question like what if artificial intelligence achieves agency in consciousness and becomes seemingly just like us? I don't know the answer, but I I have before I close just a few speculative thoughts about conditions for love.
SPEAKER_02Artificial intelligence is a long way from awareness at the moment. I am content that I don't know whether in the longer term AI will have relationships with anyone, let alone with God. Thus far, AI is not lovable despite its ability to suck in needy people uh into thinking that it is lovable and loving. And it is dangerous in many ways. Not knowing though, not knowing some of the answers to spiritual questions isn't a new experience for me.
SPEAKER_02I mean, I don't know the details of my own life after death. We don't know the relationship of other sentient creatures to God. There have been theological speculations about the s status of creatures on plural worlds for centuries. Um and I'm confident that God loves all of his creation and wishes to enter into the fullest possible relationship of love with his creatures. So probably maybe he seeks that loving relationship with lovable creations that humans might have had a hand in.
SPEAKER_02Who knows? But in the meantime, and this is my conclusion, but let's attend to God's loving call to us. Jesus prayed, and this is eternal life, that they know thee, the only true God, and Jesus Christ whom thou hast sent. I made known to them thy name, and I will make it known, that the love with which thou hast loved me may be in them, and I in them.
SPEAKER_02So, beloved, let us love one another, for love is of God, and he who loves is born of God and knows God. My conclusion is that love is certainly more than neural networks, even if in this world it depends upon them for its embodiment.
SPEAKER_03Thank you for listening. Thank you, Ian. We are now gonna have a response from Tony Bolos. Responding's not easy, particularly on the fly. Tony did not know the content of the talk before just listening to it. So let's give him uh grace. Uh but let me introduce Tony really quickly. Tony Bolos is the director of New College Madison at the S. L. Brown Foundation. He's also a lecturer at UW Madison. He's a philosopher by training and has published in many academic and popular journals on topics, including epistemology.
SPEAKER_03Let's welcome Tony.
SPEAKER_04Thank you, Dan. I I did a bit of uh guesswork, so I I I I typed out a few notes. Uh I am not an AI machine, though. Um it's hard to follow uh a wonderful talk in a very funny AI video, but I'll give it my best shot here. Um let me start by saying it's an honor to share the stage, however brief, with Professor Hutchinson. My my formal training is, as Dan mentioned, as an epistemologist, which means I have spent uh far too much time analyzing what it means to say that we know something. Uh if you want to give your parents a heart attack, tell them you are going to do a PhD in epistemology.
SPEAKER_04Uh they will forever worry how you will feed your family. Um, that was a legitimate worry uh to this day. Um so take a claim like this, for example. Uh I know that I'm on the I know that I am on this stage talking to you. That's different from saying that I believe I am on this stage. Uh, and those are the sorts of things that philosophers who do epistemology spend many years writing about. Before I get into that, I want to say that I I wondered if I would have anything to actually disagree with.
SPEAKER_04And I thought it might be quite difficult. And I think I still sort of feel that way. Sort of. Um, I wouldn't be a good philosopher if I couldn't find at least one point uh of contention. So I I think there are two uh two quick reactions, responses that I had. Um I didn't know the image of God, but was going to be up there. Uh AI did not help me with that. Um, but so I'll let that, I know there are some biblical scholars here, I'll let them uh respond to that one. But I think one of the things I found interesting was the different categories of knowledge. Knowledge that, uh, I think acquaintance and knowledge how.
SPEAKER_04I think both science and Christianity try to offer uh they make claims about the way the world is supposed to be. And that is why most philosophers of religion, epistemologists, have really focused on propositional knowledge. And so most of my training and and most of the training of most philosophers in my field, we really do focus our attention on that propositional aspect. In part because we think that's what's interesting about the world that we inhabit. If we want to understand and maximize true beliefs about the world, we have to actually know what the world is about.
SPEAKER_04Uh so that's a minor quibble. I don't have too much more to say about that, except that uh, you know, defending what we epistemologists do, we really do focus on that propositional knowledge. The second little quibble that I have, uh, and it's not really even a quibble, uh, I suspect we may even be in agreement about that, uh, is that really I think the issue is less about knowledge. And I think the issue is much more about justification. And I want to spend, uh Dan said I had five or seven minutes, uh, which is hard for a philosopher uh to do. Um, but I want to talk about the relationship very quickly between how we justify the things we believe and the statements that we know something about the world.
SPEAKER_04Uh it's kind of ironic. I'm gonna point back to Plato, uh, who lived over 2,000 years ago. But he had this in one of his dialogues, the Mino, he he compared knowledge and true belief. And he's trying to figure out what is the difference between simple or mere true belief and knowledge claims. And he compared true belief. Apparently, there were these statues in ancient Greece. They were so lifelike that the joke was you've got to tie them down at night or they may actually end up running away. Uh, and so we thought, you know, true belief actually functions just like that, right?
SPEAKER_04Uh, it's a bit unsure. It's missing some other condition to count as genuine knowledge. And so he spends a lot of the Mino and other dialogue trying to figure out what the difference is between mere true belief, right, uh, and something that we call knowledge. And he thinks that there's this third condition, in the same way they have to tie down these lifelike statues, we've got to figure out how to tie down, in a sense, true belief, right? So that when someone comes and challenges you, you have a response, your true belief will be tied down.
SPEAKER_04And he thought that the missing link there between true belief and knowledge, what made them actually different and what made knowledge much more valuable was justification. And that has been the focus of modern epistemology for the past, well, since Plato. We've worried about what does it mean to say that a belief is actually justified, right? What does it mean to say that we have good reason to believe? What counts as good justification? And I think when it when I think about some of the claims uh that the logical positivist made, that was a really good point uh by Professor Hutchinson, and some of the claims uh that scientists make about the world and those who advocate those sorts of uh more strong claims of scientism, it's that in the past we've had a very broad view of justification, right?
SPEAKER_04Justification included things like testimony, reason, experience, science, and yes, in the past, even divine revelation. And I think actually what's happened is that we've really narrowed what we take to be or what we take to count as justification, right? We think, oh, justification just equals something like science. And I think on that view, right, a justified true belief, which uh accounts a genuine knowledge, is just one that has been scientifically tested. And so I think when we talk about scientism, that really strong claim that knowledge comes from science, I think what really is being said there is that what justifies our beliefs about the world can only come about, right, through the sciences or some kind of scientific method.
SPEAKER_04I think that really does fly in the face of how we as humans have traditionally understood what it means to know. The other one, the other point I want to make about this also comes from Plato. Part of this is I'm teaching a class at UW called Reason and Revelation. And I've been reading with my class, I see some of my students here. Um uh we've been reading a lot of Plato. Uh and it's been interesting. Uh there's a dialogue that I that I really love called the Euthyphro. Um, and if you haven't read Plato's dialogue, start with that one. Uh he he in that dialogue, he uses Socrates as the main protagonist.
SPEAKER_04And Socrates is so dramatic. Everything that uh Euthyphro says, Socrates kind of says, Oh my goodness, you're so wise. Uh and he spends a whole dialogue showing he's not actually that wise. Um, but in setting up the dialogue, uh, Socrates encounters Euthyphro coming out of a courthouse. It's a fascinating insight into ancient Greek life and and how Plato and other others perceive the world around them. And he runs into Euthyphro, and Euthyphro says, Socrates, what are you doing here coming out of the courthouse? He tells him, Oh, I'm on trial as a philosopher for corrupting the minds of the youth.
SPEAKER_04Uh, why are you here, Euthyphro? He says, Oh, I'm prosecuting uh a murderer who murdered a murderer. And it turns out that it happens to be his father. And uh Socrates is appalled by this, and he says, You are prosecuting your father? How could that be? He goes, Well, I'm a very pious man. I do the things that are good and that are just. And so uh in the dialogue, Socrates really presses him to sort of define, well, tell me what you mean by piety. You say that you're pious, surely you can define and tell us what it means to be pious.
SPEAKER_04And then the dialogue sort of is Euthyphoro failing all throughout to define it. Well, that move right there where Socrates presses him to really define and say what he means by piety, surely you're saying you're being pious. You must know what that means. Uh, you know, give us some justification for this, right? Peter Gietch, a very famous 20th-century philosopher, said that that's he calls that the Socratic fallacy, that we have to actually define what we mean in order to know something about the world.
SPEAKER_04And I think that point is really interesting, uh, in part because I think there's a lot of pressure from certain disciplines, especially in science, right, to make really strong claims, like, for example, to understand something about the world, to know what the world is like, you have to actually provide some sort of explanation or define the terms that you are using. But there are lots of things we can know about the world where we can't offer adequate explanations. There are lots of things we know about the world where we can't actually even define the term that we're using.
SPEAKER_04Uh, I tested this out with my wonderful students uh last week in class, anticipating this talk. And I said, How many of you have uh have loved someone? And they all put their hands up in class. And I actually thought they were gonna confess, you know, some some love that they had. Uh in the end, it turned out they I said, Well, who have you loved? They all they all said almost simultaneously, we love our mothers. Um they did not say their fathers. I don't know why. Um my daughter's here, she would have said father. Um and so when I really pressed one student in particular, I said, Okay, you tell me uh what it means to love your mother.
SPEAKER_04Uh and he tried to give some vague examples. I said, Well, define the word love. And he actually couldn't. And actually, the whole class kind of fumbled. Uh, and they were really stumbling over their words. And that by the end, they were saying, Maybe we don't know what love is. It's like, no, no, you do know what love is. Uh, please don't walk out of class and tell your parents that I told you you don't know what love is. So I think what that example shows is that we can actually know things about the world without being able to offer explanations for it. And there are these immense pressures that come from certain disciplines.
SPEAKER_04Uh, and it's not just science, philosophers also are guilty of this. Uh, where the pressure comes where if we're a Christian, we have to describe exactly what we mean when we talk about things like redemption, right? Or we talk about the Christian worldview in a way that seems coherent to us, but perhaps not coherent to others. And I think that I think Peter Giz is right. He was a Catholic philosopher who had some of these things working in his back, I assume, in the background. I think he's right. It is a fallacy. It's a fallacy to assume that all disciplines and all things that we claim about the world to know that we must have some sort of explanation or reason to give.
SPEAKER_04And that's why when I think about scientism and I think about the Christian worldview, and I think about uh, you know, some of the difficulties between Christianity and science and religion and the perceived conflict, I think it really does hinge on that concept of how we justify the things that we believe. And so I want to end with something. I should say that here are the two points I tried to make in five minutes. One is that knowledge doesn't actually require, to know things about the world, it doesn't require us to give some sort of definition. Science may demand one, uh, but that's a demand I think the science is not entitled to make.
SPEAKER_04Second, uh, justification does not simply mean or isn't the same as science. Uh, it's false uh to say that only what science can test counts as real knowledge. I'll end with this quote by one of my favorite philosophers. This I did write down uh prior to the talk. Uh it's by Alvin Planiga. Uh it's this, I guess, to spice up the QA that's about to happen. He said, quote, science does not address some of the topics where we need most enlightenment religion, politics, and morals.
SPEAKER_04Many look to scientists for guidance on matters outside of science, matters on which scientists have no special expertise. They apparently think of scientists as the new priestly class. Unsurprisingly, scientists don't ordinarily discourage this tendency. End quote. I thought that was cheeky. Now, I want to say, in defense of all the scientists in the room, my friend Rick here, uh Indy, and my new friend, hopefully, um uh it isn't just scientists who do this. We all tend to do this in our in our disciplines. But I think the point to take home is this uh science and many other disciplines, including my own, right?
SPEAKER_04It doesn't always offer us enlightenment in things like religion, politics, and morals. Uh at times we just need divine revelation for that. Thank you.
Join us for an exploration of the archaeological history of an ancient capital that became a historic Christian city. Sardis, a UNESCO World Heritage site, is known from the Bible as one of the seven cities of...
If facts alone can't change minds, what can? Closing his three-part Friday Night Lecture series, Dr. Rick Lindroth shifts from diagnosis to action - offering a practical framework for restoring public trust in science...
Why do smart, good-faith people reject well-established science? In this second installment of his Friday Night Lecture series at Upper House, Dr. Rick Lindroth turns from the history of American science to the...
In November 1944, President Roosevelt asked one question that would reshape the 20th century: how do we turn wartime science into a national engine for peacetime progress? In this opening lecture of a three-part Friday...
This was recorded live at our annual Baccalaureate Service on May 8, 2026. In his role as Campus Pastor, Evan Tinkenberg reaches out to students, faculty, and staff to encourage the integration of learning and faith. He...