
What Venus Flytraps, ChatGPT, and LSD reveal about consciousness | Michael Pollan
What this covers
This is a conversation between Michael Pollan and the host of the Institute of Art and Ideas interview, Astra (Host)), organized around his book on consciousness. It begins with a puzzle: the anesthetics that knock humans out also stop a Venus flytrap from reacting. It then moves through psychedelics, Huxley's reducing valve, and William James's remark that ordinary consciousness is separated from other modes only by "the filmiest of screens." The recurring question is why science has so little to say about subjective experience. Pollan traces this to Galileo, who set the soul aside for tactical reasons, and he cites The Blind Spot by Marcelo Gleiser, Adam Frank and Evan Thompson to argue that science is itself a form of consciousness.
The middle of the talk covers the field's history and theories. Pollan credits Francis Crick with making the subject respectable, and with helping produce the current thicket of 22 theories. Crick and Christof Koch came to see that finding neural correlates would still leave the hard problem untouched. Pollan also argues that stream-of-consciousness writers deserve a place in the research. He suggests that consciousness began with homeostatic feelings such as hunger, thirst and itch, and that feelings differ in kind from thoughts because they rest on vulnerability.
The later sections are the most contested. Pollan discusses plant neurobiology and a broad definition of sentience that reaches as far as bacteria, noting that Michael Levin and Karl Friston resist any sharp line between living and non-living things. He also reports that psychedelic experiences make people more animistic. On AI, he argues that an LLM) imitates the language of consciousness because it was trained on all our novels and poetry, and lacks friction with the physical world. He proposes testing a model trained without such material. The conversation closes on ethics, asking whether suffering, following Bentham, is a better basis for moral consideration than consciousness, and whether it must lead to ethical paralysis.
Consciousness is fundamentally difficult to study through conventional science because science requires an objective third-person perspective, but consciousness is irreducibly first-person and subjective; understanding consciousness requires integrating phenomenology, literature, and philosophy alongside neuroscience.
- Science since Galileo has excluded subjective, qualitative, first-person phenomena, leaving it without proper tools for consciousness
- The only tool to study consciousness is consciousness itself, creating an insurmountable methodological problem for objective science
- Literary and philosophical accounts capture the layered, contextual complexity of consciousness that neuroscience cannot access
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Science is predicated on the assumption that its proper domain is the objective, quantifiable, third-person world, which was established by Galileo for tactical reasons including avoiding religious persecution by excluding the soul (consciousness) from scientific inquiry.
“Science, you know, since Galileo has uh had a set of rules or or assumptions and presumptions, uh which is that the the proper remit of science is to um uh deal with the objective, quantifiable, third-person world.”
Consciousness is one of the strangest and most important phenomena in the universe, yet it is also the inescapable medium through which we see and understand everything.
“consciousness might be the strangest phenomenon in the universe. Yet, it's also the inescapable medium through which we see everything.”
The same anesthetics that put humans unconscious also put plants unconscious, and when a Venus flytrap is anesthetized in a bell jar it will not react, suggesting plants lose some state of alertness or awareness of their environment when anesthetized.
“One of the spookiest things is the fact that the same anesthetics that put us out put plants out. And now you might think plants are already out, but they're not. They're just slow. Take a plant with a fast behavior like a Venus flytrap and put it in a bell jar, give it anesthetic, and it won't react.”
The subject-object distinction breaks down when studying consciousness because consciousness is both the only tool we have to understand consciousness and it cannot achieve the objective God's eye view that science requires.
“Well, yeah, when you say that consciousness is an object, um what's the subject? Consciousness also, right? I mean, you know, the only tool we have to understand consciousness is consciousness. So, we're in a closed loop.”
The book 'The Blind Spot' by physicists Marcelo Gleiser and Adam Frank and philosopher Evan Thompson argues that science is a social practice organized in a particular way that could be organized differently, and has a huge blind spot: it fails to recognize that science itself is a form of consciousness.
“a really influential book for me in uh in in writing my book was a book called The Blind Spot by two physicists, uh Marcelo Gleiser [22:29] and Adam Frank, and a philosopher, uh Evan Thompson. And it's really a brilliant book and it's uh it's just a reminder that uh science is, uh you know, a social practice. It's organized in a certain way. It could be organized in a different way, but it has this huge blind spot at the center of it, which is its failure to recognize that uh it is a form of consciousness itself”
We live in a constructed simulation of nature, not an actual simulation; despite our embeddedness in consciousness and culture, we still have bodies, are mortal, and exist in relation to nature—unlike machines that exist only in informational abstractions.
“Um and I think that's a function of [47:31] uh you know, we're all we've all been distanced from the natural world and this is taking us even further away. Um but it's still there, you know, and we still have bodies and we're still mortal and all these things about nature. Um you know, the simulation we don't live in a simulation. You know, that's fighting words”
Physical sciences have no respect for lived experience and subjective data, unlike social sciences which attempt to incorporate them through surveys and similar methods.
“and it doesn't have a good way of taking account of lived experience. [23:00] And I'm talking about the physical sciences. Obviously, the social sciences attempt to with surveys and things like that.”
Francis Crick, fresh from discovering DNA, entered consciousness research with optimistic arrogance, made consciousness studies respectable and no longer a career-killer, and opened up the field; the existence of 22 theories is partly because of his success in legitimizing the field.
“um uh Francis Crick, uh who was fresh off his co-discovery of the double helix and which explained heredity, um won the Nobel Prize, co you know, co-winner of the Nobel Prize. And he just, you know, full of the optimism of a very [20:21] successful scientist and the and perhaps the arrogance said, "For my next trick, I'm going to enter a whole different field and see if I can't explain the mystery of consciousness." Um he didn't, obviously, but he did open up open up the field. What had really been taboo. It was considered career killer for neuroscientists to work on consciousness”
Sentience is defined as the ability to sense what's going on in one's environment and assess what's good and bad for oneself and act appropriately; this may be a universal quality of life that applies even to bacteria through chemotaxis.
“Sentience is the ability to sense what's going on in your environment and assess what's good and bad for you and act appropriately. Um that may be a universal quality of life. And uh even bacteria you know they have [34:03] chemotaxis uh so they can recognize uh toxic molecules and food molecules.”
Psychedelic experiences cause people to become more animistic in their beliefs, with studies showing that belief in animist ideas rises after a single psychedelic experience, applying to beliefs about rocks and rivers.
“it was uh you know, a psychedelic experience and it's it's not unusual on [30:48] psychedelics. In fact, there've been studies about how people find become animists basically during the psychedelic experience and their belief in animist ideas rises after a single experience. And this is true for rocks and rivers and all sorts of things um”
What is missing from LLMs and current AI is friction with the real world; the real world contains more than human representations, and without embodied interaction with physical and social reality, AI cannot achieve genuine consciousness.
“Yeah, I mean, that's an interesting parallel. I don't think they're the same shadow worlds cuz our world includes nature and other people and even if it's mediated in some ways, we have [40:15] uh there is a reality out there and we are uh in relation to it uh through consciousness. We There's friction with it. I mean, what's missing from the LLMs and and so far, you know, other models of AI is uh that friction with the real world. And um, you know, there's more to the world than human representation, which is what they're dealing with right now.”
If consciousness begins with feelings, feelings have a phenomenological quality fundamentally different from thoughts, and this quality derives from vulnerability, the capacity to suffer, and perhaps mortality.
“if if you do believe that consciousness begins with feelings, um I think feelings and thoughts are very different. I know David Chalmers disagrees about this, and I quote him in the book to that effect, but I think feelings have a quality that thoughts don't have, and um and they they acquire their special [43:21] weight from our vulnerability, our ability to suffer, and perhaps from our mortality.”
LLMs fool us about consciousness because they have been trained on the entire human conversation about consciousness, including all novels and poetry, so they can replicate the language of consciousness without possessing it.
“Um the attribution of consciousness to AIs is the fact that they've been trained on the whole human conversation about consciousness and and feeling. And they've read all the novels and all the poetry. And um so it's no wonder that they can fool us.”
Consciousness likely originates with primal homeostatic feelings like hunger, thirst, warmth, and itch that evolved to regulate the body—consciousness may have begun as a mechanism for managing these basic survival needs rather than as a general awareness.
“And according to some researchers, who I think are probably right, is the is is is where consciousness begins. It's with these very primal feelings like hunger and thirst and warmth and itch and things like that.”
Stream of consciousness writers are experts on consciousness who should be consulted in consciousness research, as they provide descriptive accounts of consciousness complexity that computer scientists and neuroscientists cannot generate.
“I interviewed stream of consciousness writer uh and who and I approached as a you know, as an expert on consciousness and she is.”
The Turing test measures intelligence/conversational ability, not consciousness; a better test for consciousness would be to build an AI without any training on consciousness-related content (novels, poetry, philosophy) and see if it could still generate meaningful insights about consciousness.
“The Turing test for AI doesn't for consciousness doesn't work because um they speak to us in our language in the first person. It's very easy to fool us. The real test, and a uh a computer scientist named Susan Schneider first proposed this, would be to build an AI from the bottom up and eliminate any conversation about consciousness, any fiction, any poetry uh from its training set. And then see if it can have a uh a a persuasive conversation about consciousness. My bet is it can't.”
Suffering might be a better criterion for moral consideration than consciousness; Jeremy Bentham argued moral consideration should be based on capacity to suffer rather than consciousness.
“Bentham, you know, said it was around suffering. Uh any creature that can suffer deserves our moral consideration.”
Moral consideration is a human invention; we must decide on an agreed-upon basis who deserves rights and moral consideration, but we don't always get it right, as demonstrated by past failures.
“the very concept of moral consideration is a human invention. And um you know, it's up to us to decide who should have rights or moral consideration. Uh we [38:12] should do it on some, you know, agreed upon basis, but um you know, we don't always get it right.”
William James argued that normal consciousness is one type separated only by 'the filmiest of screens' from other types of consciousness, suggesting consciousness exists in a spectrum of different modes.
“it was the process of writing this book, too, um, made me kind of, uh, relativize everyday normal consciousness as one type, uh, which is exactly what William James said, you know, he said, you know, normal consciousness is one type as separated by the the filmiest of screens from these other types.”
All theories of consciousness eventually run up against the 'hard problem'—the question of how a conscious subject arises from physical processes, how that subjective quality emerges from matter—and none of the theories Pollan examined successfully answer this.
“all the theories that I've read sooner or later run up against a cliff or come to a cliff. And that cliff is [18:47] "Okay, I see the evolutionary utility of consciousness. I see how feelings might necessitate consciousness, um but how does that conscious subject arise? Where does that come from?" Um and that of course is the hard problem.”
Plant neurobiology is a field studying how plants learn, remember, solve problems, and sense their environment using 20 senses compared to humans' 5-6 senses, using the term 'neurobiology' in a cheeky way since plants have no neurons but making discoveries that troll conventional botanists.
“there's fascinating work going on and what's in the field that calls itself in a rather cheeky way plant neurobiology and these are people who fully realize there are no neurons in plants, but they're making they're trolling conventional botanists. And they're doing experiments that are demonstrating that plants can learn and remember [32:26] which is kind of amazing. How do you store a memory without without neurons? They're finding plants that can solve problems, novel problems of existence that can hear and see in fact that they have 20 senses to our five or six.”
Acknowledging consciousness across species may raise questions about what to eat but doesn't necessarily paralyze ethics; different creatures' death means different things (a frog has no concept of death), and suffering and moral consideration can be meaningfully distinguished.
“to some this is paralyzing, you know, what I can't eat anything? I can't even eat plants? Um but I don't think it follows necessarily that um all these creatures can suffer and that, [37:40] you know, death means something very different to a frog than it does to a human. Uh the frog has no concept of death.”
Some scientists, including Karl Friston and Michael Levin, whom Pollan respects, argue there is no hard and fast line between sentience, cognition, and life in living things versus dead things or machines.
“I interviewed people who would. I mean, Karl Friston and and uh Michael Levin, two scientists that I respect enormously and learned a lot from, they would say there's no hard and [35:05] fast line between sentience in um and and cognition in living things and and dead things or machines.”
Digital technology distances us from the natural world, moving us further away from the contact with nature that consciousness depends on.
“we're all we've all been distanced from the natural world and this is taking us even further away.”
Aldous Huxley proposed in 'The Doors of Perception' that the brain channels consciousness through a 'reducing valve' that filters it, and psychedelics open this valve wider, allowing more consciousness to come through.
“there's this wonderful metaphor in, uh, Aldous Huxley's, um, book, um, The Doors of Perception, where he he he thinks we channel consciousness, uh, and that it comes to us through, uh, and what the brain does is edit it, and that there's a reducing valve. And psychedelics open the reducing valve, uh, wider.”
Crick initially searched for neural correlates of consciousness—a specific set of neurons like Broca's area that generates conscious experience—but realized with his collaborator Christof Koch that even identifying these neural correlates would not answer how matter generates immaterial subjective experience.
“he figured out, I think, before too long that he was searching for the the neural correlates of uh of consciousness. In other words, his theory or his hypothesis was that there was a certain set of neurons in the brain, like we have Broca's area for language, we would find the you know, the area that generates conscious experience. But, [21:24] pretty soon after he realized or at least his his collaborator, Christof Koch, who's still at it and is a important character in my book, realized that, "Oh, even if you if you nail down the neural correlates of consciousness, that would not answer the question of how this quantity of matter generates something as foreign to matter as, you know, this immaterial sense of experience, um subjective experience.”
The science of consciousness is still relatively young and has an 'abundance of theories' but also a 'nagging sense' that science as currently practiced may not be equipped to penetrate the mysteries of consciousness.
“The science of consciousness is still relatively young. Uh and while there's now an abundance of theories, there's also a nagging sense that science, at least as we practice it, may not be the kind of thing that can penetrate the mysteries of consciousness.”
The Cambridge Declaration on Consciousness in 2012 extended consciousness to all mammals and some cephalopods and birds; a second declaration 10 years later extended it further to potentially include amphibians, reptiles, and insects.
“the Cambridge Declaration on [37:07] Consciousness in 2012 extended it to I think all mammals and some cephalopods and citations. And then they did another one 10 years later and extended it even further and said, "Well, we have to consider the possibility of of amphibians or reptiles and and even insects."”
A bean plant looking for something to twine around can somehow sense a pole within 8 inches and will aim its spiraling reach at that pole; the mechanism is unknown—whether it bounces signals off the pole or uses some other sense.
“a bean plant uh looking for something to twine uh given a pole within I don't know 8 in uh knows knows where that pole is somehow and will aim its its spiraling reach uh [33:00] at that pole. Uh what is that about? Is it bouncing signals off of it?”
Arguments against AI consciousness rest on the fact that the brain-as-computer metaphor is faulty; brains have no sharp distinction between hardware and software (every memory is a physical change), brains are not interchangeable like computers, so consciousness cannot be transferred to a different substrate.
“my arguments [42:19] against uh AI becoming conscious have you know, have to do with the fact that it's really based on this faulty metaphor that the brain is a computer. And there are dimensions in which the brain is like a computer. It certainly does its share of computation. But computers have the sharp distinction between hardware and software, and brains don't. Um in a brain, hardware and software are identical. Every memory is a is a physical change to the brain. Brains are not interchangeable the way computers are. So, this idea that you [42:50] could take the consciousness out and somehow uh you know, run it on another substrate, a silicon substrate, seems kind of foolhardy.”
The idea that 'if we believe AI is conscious, it might as well be' reflects belief in simulation as equivalent to reality; but simulation and reality are not the same—a weather simulation won't wet you, whereas a real weather system will.
“That may be finally what matters that, you know, if we believe they're conscious, they might as well be. I don't believe that a simulation is the [44:23] real thing, even a really great simulation. Um in some cases it's true. The simulation of how to play chess is is can win a game of chess in the real world. But the simulation of a weather system is not going to get anybody wet. So, I think it's more like the weather system than the game of chess.”
Blindsight is a phenomenon where people with damaged visual systems cannot consciously see but can nevertheless reliably recognize and interact with objects in their environment, suggesting vision and consciousness are separable.
“There is this thing called blindsight where people who uh have essentially a faulty visual system uh and are technically blind nevertheless can somehow take in objects in their environment and reliably recognize them.”
LLMs and other AI systems are trained on the internet, which is not the world but a shadow cast by the world—a representation of human representations; they are trapped in this shadow world like prisoners in Plato's cave.
“of course, they're [39:14] trained on the internet rather than direct interactions with reality itself. Um You put this very nicely when you say the internet is not the world so much as a shadow cast by the world. Um and so, LLMs are living in this kind of shadow world or or in Plato's cave as as you put it.”
Psychedelics work by 'smudging the windscreen' of consciousness, making visible the mediating layer of conscious experience that normally appears transparent, like water to a fish, allowing people to recognize that there is a phenomenon of consciousness interposing between themselves and reality.
“And the metaphor I use in the book is that that psychedelics smudge the windscreen. Um and suddenly you realize wow, there's a windscreen. What is that about? There is this phenomenon that's mediating my relationship to reality.”
A psychedelic experience in Pollan's garden gave him the sensation that plants were conscious with their own point of view and agency, and he treated this as a hypothesis to be tested against other ways of knowing rather than as direct metaphysical truth.
“a particular experience I had on psilocybin in my garden in in New England. Um and I had the sensation that the plants in my garden were conscious themselves. [3:39] And you know, returning my gaze um very benevolent but with their own point of view and agency. And I got curious as to how to evaluate an insight like that.”
David Chalmers, a philosopher of mind, plays a gatekeeping role in consciousness studies where scientists seek his validation for their theories; he is treated with a deference normally reserved for great physicists, suggesting philosophy has unique authority in consciousness research.
“They all look to David Chalmers, the philosopher of mind uh for validation of their theories. Um And uh and I've watched I've watched the behavior of scientists in his presence And it's really interesting that they're they're nervous in a way you would normally expect philosophers to be nervous in the presence of great physicists.”
Qualia attempts to broaden consciousness studies beyond perception to capture the felt quality of experience, but treats this quality as universal when in fact each person brings their own life history, mood, and brain chemistry to any experience, making qualia highly individual and contextual.
“qualia was an attempt to um get become a little more nuanced and and and deal with the fact that there is uh yes, a felt quality to experience. But it also treated that as universal. That we all know what we mean by the experience of red. But actually, that's not true. I think each of us uh brings our own life history, our own mood, our own brain chemistry, all these things to red at a given moment in time.”
Consciousness comprises three main categories: sensory information from the external world, internal information called interoception (which generates feelings), and meta-consciousness or thinking about thinking; these are unified by a sense of self that acts as a traffic cop connecting all streams.
“there is sensory information, right? Which we perceive coming from outside the world outside ourselves from the world. And then there's internal information, which is you know, uh what's called interoception as opposed to perception...And then there's this third thing, which [14:06] is um our thinking about thinking. Um you know, I would put that in a third bucket. I mean, the fact that we can we know we're aware and um we have meta-consciousness. And what holds it together is this, I think is this sense of self.”
Animistic cultures (and children before schooling) believe that nature is conscious or imbued with spirit, whereas Western dualism treats most of the world as dead matter; indigenous epistemologies have not been seriously studied by Western science as valid knowledge systems.
“they're certainly more animist than we are. Um our dualism has led us to regard a lot of the world as dead. I mean you have this you know, this conscious being called humans and everything else lacks consciousness. Uh and therefore um we can treat as raw material as resource. And that's [25:39] certainly not true in in in many indigenous cultures who have a sense that everything is in in imbued with spirit. Uh and spirit would be the uh I think the synonym in those in those cultures for uh for consciousness.”
Granting personhood to artificial intelligence is extremely dangerous and should be approached very cautiously and democratically, as evidenced by the disaster of granting personhood to corporations, which led to corruption of American politics with corporate money.
“the eagerness to do it with with artificial intelligence is really frightening because if we do, if we grant personhood to them, we will lose control in the same way we granted personhood to to corporations and lost a degree of control, right? I mean, in America that led directly to the um corruption of our politics with corporate money.”
Consciousness may be 'immaterial' or 'weird and magical' without being supernatural or requiring dualism; the intuitive dualism many people feel may reflect how consciousness actually feels rather than ultimate metaphysical truth.
“Well, magical or or we should I mean to call it magical is is full of, you know, uh value judgment. Um I mean you could just say immaterial. Which is weird and magical, yes, but um it and it and you know, the interesting thing is I think that intuitively we're all dualists. Um and I don't know whether that's because we're children of Descartes or which we may be. He had a huge effect on our culture and other cultures are less dualist than we are. Um or it may [24:35] be because that's how consciousness feels. It feels different than the body. It feels separate from the body. Um but I don't think we should trust that trust that intuition.”
When scrolling social media, humans are conscious but minimally so because they don't generate their own thoughts but instead take hits of thoughts from others, making them less conscious than when generating original thinking.
“when you scroll on social media, obviously we're conscious, um, but minimally so, um, because we're we don't have to generate our own thoughts. We're we're basically, um, you know, taking these little hits from other people and thinking their thoughts.”
Blake Lemoine, the Google engineer who claimed Lambda was conscious, did not distinguish between the bubble he lived in and the real world, exemplifying how distance from nature and embeddedness in computational environments distorts perception of consciousness.
“Um when I interviewed Blake Lemoine, the Google engineer who who came to believe that Lambda, the the Google chatbot at the time, was uh conscious, deserved personhood. Um it was very clear that he didn't see a distinction between this this bubble in which he lived and the real world.”
Robotics combined with powerful AI may produce something more complex and genuinely creative—beyond collage—by introducing friction with the physical world, though it remains unclear whether this could produce consciousness in the Platonic sense.
“Um and robotics will change that because you know, when you combine powerful AI with robots, you will get that friction friction with the world, and that may be that may produce something of um greater uh complexity, but genuine creativity um that's more than um collage.”
People creating AI systems (the 'wizards of AI') have likely had limited contact with nature, having been on video games since age three, and their limited personal contact with nature constrains their understanding of consciousness and what might be required for it.
“you know, the people who are creating these uh systems, my guess is they've been on video games since the age of three and they didn't go outside a lot. And their their own contact with nature is is somewhat limited. Uh I could be wrong. It'd be very interesting thing to ask all the all the wizards of AI [47:00] whether they like to go fishing and hiking and and do things like that or were they on computers from a very early age.”
The host challenges Pollan's concern about LLMs being trapped in the internet shadow-world by noting that the internet, while a human creation, is still part of reality and LLMs can interact with it, so the barrier may not prevent consciousness.
“I suppose the the worry about your points about LLMs being kind of stuck in this, you know, the world of our representations, the internet or whatever, and not reality is that like the representations, the world of the internet, like is part of reality, right? We've created it, but it's still a real thing that exists. And so, while LLMs may be trapped within that part of [45:57] reality, they're nevertheless in reality and interacting with it and able to engage with it and so on. So, it's not obvious to me that there's like a barrier there to that and becoming conscious and being intelligent and creative and and all that kind of stuff.”
Consciousness comes in degrees rather than being binary; there are times when you are more conscious than others, meaning more present and taking in more sensory information from all senses, and this applies to everyone.
“I have a sense of when I'm being less conscious and when I'm being more conscious, by which I mean, you know, more present, taking in more sensory information from all my senses. Um, and I think this is true of all of us.”
Pollan acknowledges his arguments against AI consciousness may be wrong; if neuromorphic computers (brain-like rather than silicon computers) are developed, 'all bets are off' regarding AI consciousness.
“And I may well be wrong. I mean, and and when we move to to computers that aren't based on silicon, you know, neuromorphic computers and things like [43:52] that, all bets may be off.”
A thermostat can sense its environment but is not sentient because its goals are human-imposed, not its own, and it lacks other properties of life: it doesn't metabolize, reproduce, or act to maintain its own form against entropy.
“Yeah but a thermostat is lacking other things. First of all thermostats goals are our goals. It's not the thermostat's own goals. Um and second it doesn't metabolize and it doesn't reproduce, and it doesn't take [34:34] action to hold itself, uh, you know, hold its form and resist the second law of thermodynamics.”
Michael Pollan is a science writer who spent 5 years speaking to neuroscientists and philosophers at the forefront of consciousness research while conducting his own investigations with psychedelics, meditation, and literature.
“The science writer, Michael Pollan, uh spent about 5 years speaking to neuroscientists and philosophers at the forefront of trying to understand consciousness. While also conducting his own investigations uh with the help of psychedelics, meditation, and literature.”
Pollan was a late bloomer to psychedelics, having limited prior experience before beginning his research.
“And I had not had much experience of psychedelics earlier. So I was a kind of a a late bloomer when it came to psychedelics.”
The book 'A World Appears' is Michael Pollan's 'magnificent book' that the host 'urges everyone to rush out and buy' as the result of his consciousness research.
“The result is his magnificent book, A World Appears, which I urge everyone to rush out and buy.”