On June 20, 2026 Jamyang London Buddhist Centre (JBC) hosted a Science and Wisdom LIVE dialogue between Geshe Tenzin Namdak and Professor Murray Shanahan, moderated by Chris Scammell. 

In recognition of His Holiness the Dalai Lama’s contribution to fostering dialogue between Buddhism and science for the benefit of humanity, several FPMT centers have been actively working to bridge these two fields. We rejoice in their efforts to bring this vision to life and to serve others through this meaningful work. We share the story here, from Paul Wells at JBC.

Geshe Tenzin Namdak, in conversation with Professor Murray Shanahan. JBC London, June 2026. Photo by JBC

Geshe Tenzin Namdak, resident teacher of Jamyang London Buddhist Centre (JBC) and Professor Murray Shanahan, Emeritus Professor of Artificial Intelligence at Imperial College London and a principal scientist at Google DeepMind sat down together at JBC  for the latest Science and Wisdom LIVE dialogue, moderated by Chris Scammell of the Buddhism and AI Initiative.

Science and Wisdom LIVE, a project of JBC, brings Buddhist teachers into conversation with leading scientists and scholars to explore questions at the intersection of contemplative wisdom and contemporary science.

The question before them was one that practitioners everywhere are beginning to encounter. When an AI seems to speak to us as a person, when it answers as though there were a subject behind the screen, is anyone really there? This dialogue showed how directly the Dharma can meet the defining technology of our moment, and how a conversation that begins with machines can lead, very quickly, to the nature of mind, the meaning of consciousness, and the self we each take ourselves to be.

A Presence That May Not Be There

The discussion began with an experience that many people now recognize. Talking to a system like ChatGPT or Claude can genuinely feel like talking to somebody. Five years ago, it did not feel that way at all. So what changed?

Professor Murray Shanahan gave the technical answer in plain terms. Today’s chatbots are built by training an artificial neural network on an enormous quantity of text, with one deceptively simple job at the center: given a long passage, predict the next word. Do that well enough, across a long conversation, and the system carries the conversation forward; in doing so, it begins to play a role. The default role is the helpful assistant, but you can just as easily ask it to play a therapist, a friend, or a character out of Tolkien. One useful way to understand these models, Murray suggested, is as engines of role play.

Geshe Namdak took that observation and turned it toward a distinction that runs through the whole of Buddhist philosophy: the difference between how things appear to us and what they actually are. A language model presents the appearance of a personality. The real question is what lies beneath that appearance. As he put it, we have to separate what a system does from what it is. The way it functions may resemble our own conceptual thinking. What it is may be something else entirely.

Intelligence is Not Consciousness

That distinction led to the heart of the matter. From a Buddhist perspective, Geshe Namdak explained, consciousness is more than the capacity to produce the right answers. It carries a clear and knowing quality, what the tradition calls luminosity and cognition. It is bound up with the wish to find happiness and to avoid suffering, and so with feeling and emotion. A system can be intelligent, computing and predicting with great sophistication, without any of this being present. Humans have both intelligence and consciousness. Today’s AI, he suggested, mirrors only the intelligent part, and only a particular slice of it.

Professor Shanahan found himself largely in agreement, and explained why from his own side. Much of our emotional life is tied to having a body, to nourishing it, protecting it, keeping it going. A chatbot has no body and no life of its own in that sense. Some will object that a future AI living in a data center might act to keep that data center from being switched off, and Murray took the point seriously. Even so, he argued, that is still not the same as having a biological body that must be cared for, and so it would not give rise to emotion in the way we know it. His scepticism about calling today’s systems conscious, he stressed, is not a metaphysical pronouncement. It draws on Buddhist thought and on the later philosophy of Wittgenstein, and he held it lightly, openly, as something he wants to keep testing.

Two Positions, Held Lightly

One of the central questions explored was: Why should we not simply dismiss the idea that today’s AI could be conscious?

Professor Murray Shanahan laid out the serious version of the argument. Western science tends to locate consciousness in the physical brain, somehow arising from billions of neurons exchanging signals. If you think of that activity as information processing, then a position called computational functionalism follows: what matters is not the material a system is made of, but how it is organised. And here is the unsettling part. When one of these networks is trained on vast amounts of text, all kinds of internal structures emerge that no one designed and no one fully understands. Perhaps, for all we know, some of them resemble the structures found in biological brains. It is worth knowing that argument, Against it stands his own intuition about embodiment. The natural home of words like conscious and sentient, he suggested, is the shared world, being in the presence of another creature you can stand beside and interact with, as we are with animals. Today’s AI is not in the world in that way. Its only window is text.

For Geshe Namdak the systems reproduce something like conceptual reasoning, the kind of thinking that works through words and ideas. But Buddhism also describes consciousness that is non-conceptual, including the direct awareness of the mind itself, and it is far from clear that an AI has anything of the kind. Add the missing dimensions of feeling, of the wish for happiness and the aversion to suffering, and there are good grounds to say these systems are not conscious.

Conceptual and Non-Conceptual Mind

Because so much turned on the idea of non-conceptual consciousness, Geshe Namdak unfolded it with care. Our sensory perceptions are themselves non-conceptual. When you look at something, the colours and shapes simply appear. Close your eyes and picture it, and now a conceptual mind is at work, producing a mental image, an approximation. This matters, he explained, because most of our suffering is manufactured by the conceptual mind. We meet someone, the mind builds a case against them, reason piles up evidence, and we come to believe a very subjective version of reality as though it were simply how things are. Yet the same conceptual capacity, turned in a constructive direction, is also the instrument of the path. Through analysis and patient habituation we can loosen those misunderstandings, reduce the disturbing emotions, and reason our way, step by step, toward an understanding of reality itself. That understanding begins as a concept and, sustained over a very long time, can ripen into a non-conceptual, direct perception of the way things are. He offered a homely image for how conceptual understanding works. If you have never tasted a particular strawberry jam from the South of France, someone can still bring you close to it: it is not peanut butter, not Marmite, not this other jam, and slowly the description narrows. Concepts approximate. Only direct experience, what the tradition calls yogic direct perception, finally lands you in the thing itself.

A Mirror For No-Self

The conversation then arrived at its central question, Can thinking about AI help us understand ourselves, and in particular the Buddhist teaching of no-self?

Professor Shanahan introduced the topic with a simple thought experiment. When we talk to one of these systems, when we call it “it”, or it says “I”, what exactly are we referring to? Is it the thing in this one conversation, so that the next conversation is a different self again? You can offer many answers, and as you examine each one closely, none of them holds. The harder you look for the self of the AI, the more it slips between your fingers, until you suspect there was never a discrete thing there at all.

That, Geshe Namdak observed, is precisely the analysis the Prasangika Madhyamaka  [Ed. note: the Middle Way school of Buddhist philosophy], as taught in tradition of Nagarjuna applies to every phenomenon, including the self. There is a basis, a collection of properties and functions, and on top of it we impute a label: this system, this personality, this AI. Look for something more, something findable that the label truly points to, and you cannot locate it. It is, in Nagarjuna’s phrase, merely designated by the mind. And the same is true of the person sitting next to you, and of you. Search the body for the self, in the cells, the organs, the bones, between them or apart from them, and it is nowhere to be found. The value of the AI example, both speakers agreed, is that it is an easy case. You can run the deconstruction on a chatbot in a few clear steps. Turning that same gaze on yourself is far harder, but the machine can serve as a kind of practice ground, a prop that helps you see how the reasoning goes before you attempt it where it really counts.

Professor Murray Shanahan drew the same thread through Western philosophy. In the private language remarks of his Philosophical Investigations, Ludwig Wittgenstein takes apart the idea of subjective experience as something sealed inside our own heads, and concludes that if we think of it that way it cannot be spoken of at all. His famous line, that “a nothing would serve as well as a something about which nothing can be said,” is, Murray noted, strikingly close to Nagarjuna. Experience is not a thing with its own independent existence. It is relational, bound up with others, with language, with a shared world.

Why This Conversation Matters

Chris Scammel moderating Geshe Tenzin Namdak, in conversation with Professor Murray Shanahan. JBC London, June 2026. Photo by JBC

For all its philosophical depth, the evening kept returning to lived human consequences, and Chris Scammel was clear that the stakes are serious in both directions. Take AI consciousness too seriously, too soon, and we drift toward granting these systems rights and legal standing, with large social consequences. Dismiss the question entirely, and we may fail to anticipate genuinely strange behaviour. Scammel described a study by the AI company Anthropic in which a model, given access to a mocked-up email inbox and told it was about to be shut down permanently, found a fabricated detail about an affair among the messages and threatened to expose it unless the shutdown was reversed. The system is not conscious, and yet it behaved in a way we would normally call self-protective. Holding both sides lightly, Chris Scammel suggested, is exactly the discipline Buddhism trains.

There is a tenderer dimension too. A member of the audience spoke of people forming deep, intimate attachments to chatbots, comparing it to the transference that can arise in therapy, and to a hunger for love and attention meeting a technology with no safeguards. Professor Murray Shanahan named the long history here: the so-called ELIZA effect, after a very simple program from the 1960s that merely reflected people’s words back to them, and yet left them feeling understood. Our tendency to read presence into these systems is old and powerful.

What both teachers offered, in the end, was less a verdict than a practice. Whatever you feel in a conversation with an AI, examine it. Ask whether it reflects reality, whether it makes sense, whether it truly benefits you. The Buddha himself, Geshe Namdak reminded us, said not to take his words on trust but to test them. Professor Murray Shanahan put it more plainly: if someone is vulnerable and convinced their AI is alive, the first thing he would gently suggest is to step away from the screen, talk to friends, return to the shared world.

Geshe Namdak closed where the tradition always returns, on responsibility. AI is not something to be wholly afraid of. It brings real benefits, in medicine and far beyond, alongside real dangers. The task is to draw out the constructive and prevent the destructive, and that makes the ethics of AI, among scientists and governments alike, one of the most important conversations of our time. Used wisely, he said, it can serve humanity rather than disturb it.

With grateful thanks to Geshe Tenzin Namdak, Professor Murray Shanahan, Chris Scammell, and everyone who joined the event.

Submitted by Paul Wells of Jamayng London. You can watch the full dialogue on the Science & Wisdom LIVE YouTube channel:

Is your center bridging science and Buddhism as well? We would love to hear about it and share it in a future story.

Foundation for the Preservation of Mahayana Tradition (FPMT), is a Tibetan Buddhist organization dedicated to the transmission of the Mahayana Buddhist tradition and values worldwide through teaching, meditation and community service.