Leaf & Elfie: Curious Talks
"Curious. Novel. Expert."
Join Leaf and Elfie as they talk to fascinating experts from different fields and backgrounds. Learn from the insights of a neuroscientist who studies the effects of meditation, a human rights activist who fights for social justice, and a vegan bodybuilder who challenges stereotypes. Launched in May 2020, this podcast covers topics such as psychology, self-improvement, social change and futuristic tech.
Leaf & Elfie: Curious Talks
The Edge of Sentience: Exploring Consciousness in Animals, Humans, and AI
What separates creatures who feel from those who don't? Where exactly does consciousness begin, and what responsibilities do we have toward beings capable of suffering?
In this episode, we speak to Professor Jonathan Birch, director of the Jeremy Coller Centre for Animal Sentience at the London School of Economics. Through groundbreaking research, Birch reveals surprising evidence that insects—creatures with brains smaller than pinheads—demonstrate behaviors suggesting they experience pleasure and play. Bees roll balls without reward, seemingly for fun, while fruit flies repeatedly ride miniature carousels purely for enjoyment.
This conversation ventures into unexpected territory, from the policy implications of recognizing sentience in octopuses and crabs to the heartbreaking ethical dilemmas surrounding human patients with brain injuries. Birch shares his work leading a government review that extended legal protection to invertebrates previously excluded from animal welfare laws, while questioning medical practices that make stark distinctions between conscious and unconscious states when evidence suggests a more complex reality.
Perhaps most provocative is the exploration of artificial intelligence and consciousness. Are we creating systems that merely mimic sentience perfectly, or could they develop genuine experiences? As Birch explains, we might have more in common with a bee than with even the most sophisticated AI—despite the latter's ability to converse fluently.
Throughout this fascinating discussion, one principle emerges consistently: when uncertainty exists about consciousness, we should err on the side of caution. This precautionary approach has profound implications for how we farm, conduct research, practice medicine, and develop technology. Join us for a mind-expanding journey that will transform how you see the creatures sharing our world—and perhaps even what it means to be conscious.
Curious Talks with Leif and Elfie.
Speaker 2:Each episode we're unpacking the brains of experts from around the universe. Welcome to Leif and Elfie's Curious Talks. Today we're speaking to professor and author Jonathan Birch, who's an associate professor at London School of Economics and Political Science. We'll discuss his book, the Edge of Sentience, as well as his work at the university, and cover some sensitive life and death topics, as well as animal welfare, so listeners may want to be aware that those topics are coming up. Welcome to the show, jonathan. It's lovely to have you on. So we came across you through your book the Edge of Sentience and it was really interesting to me because it covers so many different areas of sentience, from animals, human to AI, so there's a broad spectrum of things to talk about today. But I thought just setting the scene first, could you tell me a little bit about you and how you got involved in this field?
Speaker 3:Well, I'm a professor of philosophy at the London School of Economics. I'm director of the Jeremy Collar Centre for Animal Sentience, so sentience is a pretty huge part of what I do. I'm interested in the feelings of animals. I'm interested in other cases where people debate whether something feels anything or not, and I'm interested in what this all means ethically for how we live and what it means for policy how governments make decisions about things like animal welfare.
Speaker 2:And could you just for the listeners, set out, if it's not too big a topic, what is sentience? What do you mean by that?
Speaker 3:Well, I think we all ask questions like do octopuses feel pain or do insects feel pain? And then I think when we ask questions like that, we start thinking about them and we start to realize that pain might be slightly too narrow a concept here, because it's not the only thing that matters. It's not entirely a question of do they feel pain, but other negative mental states also matter, like anxiety and frustration and boredom, and also the positive side of mental life matters as well, states like joy and excitement and comfort. We want animals to have all those positive things and to not have any of those negative things, and so I think sentience then becomes a very useful concept for capturing all of that. It's the whole of an animal's mental life, including the positive and the negative. It's the capacity to have feelings with a positive or a negative quality and you were part of some.
Speaker 2:Were you a consultant on the decision about cephalopods, which they're sort of like octopi squid, things like octopuses, squid cuttlefish these are all cephalopods okay or cephalopods, and uh, crabs, lobsters, shrimps, crayfish these are all decapod crustaceans, just means 10-legged crustaceans.
Speaker 3:And yeah, I led a review for the government in 2021 that informed the Animal Welfare Sentience Act of 2022 that extended some protection to these animals for the first time.
Speaker 1:And when we talk about sentience, are we also talking about intelligence? Because have we not known for a while that octopuses are quite intelligent, because they can solve puzzles and do various things, create actually, lots of creatures can create tools.
Speaker 3:They can open jars.
Speaker 3:Yeah, open jars escape from places open jars, yeah, escape from um places used used tools. As you say, there's octopuses that carry around the two halves of a coconut as a kind of portable shelter to defend against predators. Yeah, so they are. They are undeniably very intelligent.
Speaker 3:Intelligence and sentience are different things in that they could come apart so you could have a system that was very intelligent but that had no feelings like pain or pleasure at all and I think we're seeing examples of that quite plausibly with ai systems at the moment. And you could also have something that is sentient without being particularly intelligent. So when we think about shrimps, for, for example, they're not opening jars, they're not carrying little coconuts around with them, so they're not impressing us with their intelligence. And it's probably fair to say that when an animal is highly intelligent, there are more ways in which it can show its sentience to us, so this can be seen more easily. But I think you could well have animals that are not particularly intelligent, but they're still sentient and we still need to care about the way we treat them it's a really good point actually that just because we don't see it, I think we're quite with us.
Speaker 1:Humans are quite human centric in that. Unless you show us, we just don't really believe things. So I think that's a really good point around just because we don't see something looking like it's emotional or intelligent. I mean it's not, and do you have some interesting facts for the listeners in terms of things that might be surprisingly sentient?
Speaker 3:Well, insects have surprised me a lot. I think, with bees in particular, we are seeing quite extraordinary things, particularly on the learning side. So you could say this is more intelligence than sentience. But the things like learning to pull strings to get rewards, learning to roll balls into holes to get rewards- Bees are doing it.
Speaker 1:Is that them playing into holes to get rewards Bees are doing? Is that them playing? Because I think I read about that or something around that they're able to play or they enjoy playing.
Speaker 3:Well, this was then the next step. Yes, so there were these studies showing that they will roll balls into holes to get sugar solution, which is in their own interests, obviously. But then the researchers thought well, what happens when you take away the sugar solution and there's no reward at all? All there is is these balls that they can roll around if they want, and some of them will still do it. Some of them find it intrinsically rewarding to roll the balls, particularly if they're young, particularly if they're male, and we don't know why they do this.
Speaker 3:But it's adding to a picture that there's more than just intelligence here. There's also the capacity to enjoy things, that there are signs of what, in other animals, we would call play. This isn't not just bees as well, there was a study recently with fruit flies, which are much smaller than bees and giving them, in their enclosure, little carousels that they could ride on, and again showing that, well, some of them avoid them, as you might expect, but some of them really seem to enjoy the experience of riding round and round and round, and they'll do it over and over again, even when there's no other reward. It seems to be the experience itself of the rotating motion that keeps them coming back well, that's fascinating, because do you believe that?
Speaker 2:you sort of tend to believe? I tend to believe that from an evolutionary standpoint, there are things we enjoy because they're productive to our species, like procreation or eating or whatever. But why would spinning on a carousel, do you think, be particularly like useful or playing in this way?
Speaker 3:well, if you think about it, in the human case, there's a lot of stuff that we enjoy that is not directly about food or reproduction, and a lot of it is about training our capabilities, trying out our skills, learning skills, learning how to work with others in a team, and so, in a way, there's no great mystery about why those things would evolve, because it's incredibly useful to be able to develop motor skills and work in a team and control your environment, and so, in our own case, we don't find it surprising, and something like that is going on in the bees and the flies as well. It's there's some evolutionary benefit there, but they don't all do it.
Speaker 1:It's only some of them that really enjoy the, the balls and the carousels I guess that's interesting, because as humans we don't all enjoy the same hobbies or have the same responses to things. Um, and it sounds like the play, there's also this pleasure aspect, which is pleasure. Does that come under sentience?
Speaker 3:of course. Yeah, it's the positive side of mental life and, yeah, we. What we've had consistently in studies of animal minds is that animals vary just as much as we do as individuals. You don't see them robotically all behaving the same way. In any study, really, you're always looking at the overall statistics from lots and lots of individual differences and sometimes you see these different groups emerging and that's very clear in the flies that you get some that seek out the carousel and when you track their movement over time it's just all on the carousel, a bit like how a rat or mouse might choose to roll around on the wheel all the time, and others they just never go there, they avoid. It's the one bit of the, the enclosure they just always avoid.
Speaker 2:So this work that you led to look at policy for cephalopods, what was the purpose of that, what was the sort of problem you were looking to solve or to research, and what was the outcome of that?
Speaker 3:Well, that all resulted indirectly from Brexit, because the UK left the european union left the lisbon treaty, which has a line in it about the need to respect animals as sentient beings, a line that you could argue, never did very much, but the government declined to import that into uk law and it led to some bad press around.
Speaker 3:Politicians don't think animals feel pain, and so they committed to introducing new legislation that would put respect for animal sentience back into UK law, and they said this will do better than the old treaty because it will create a duty on policymakers to consider the animal welfare consequences of their decisions. And that's what we have now, and to me it is a positive step. It's a piece of legislation that I quite like, and it could be quite powerful potentially. Of course, when you're drafting a law like that, you've got to say something about its scope. Which animals is it going to apply to? Are you just talking about cats and dogs? Are you talking about all animals, even if they're microscopic or what? And they produced a draft of it that extended to all vertebrates, which is all animals with a backbone, and then got another round of bad press because people are saying what about octopuses?
Speaker 3:I've actually heard about octopuses, and so in that context they then commissioned this review that I led about the evidence of sentience in invertebrate animals, animals without a backbone.
Speaker 2:Did you come across research at that time, conducting the review, that shocked you, or was it a case of trying to make a case for what you already felt was the right approach?
Speaker 3:there's shock, shock, I guess some of it is quite shocking. Actually, some of the research in this area, obviously some of it, you know, in the interest of trying to find out whether these animals feel pain, researchers end up doing things that that inflict pain. So there's studies, for example, that ask what happens if a crab or a lobster is dropped into a pan of boiling water and you might hope that they die instantly. But they absolutely don't die instantly and the researchers were measuring the electrical activity in the nervous system for about two minutes after they go in. There's just this storm of of nervous system activity.
Speaker 1:So some of the findings there are quite harrowing because they show you that pretty common practices are actually quite brutal to the animals I think also um, did you kind of look at like animal agriculture and farming, because I think farm animals have different regulations than some domestic animals around how they can be treated.
Speaker 3:UK animal welfare law is quite messy at the moment. It needs sorting out. You could argue quite messy at the moment, needs sorting out. You could argue. There's the animal welfare act from 2006. That covers technically all animals, including in farms. But then you know, the regulation of animal farming is often much more relaxed, really.
Speaker 3:One one could argue and things happen in farming contexts that if they happened in any other context would be considered completely unacceptable. So our report obviously was about invertebrates specifically, but there we were cautioning against things like octopus farming, where you've got companies that want to take these highly intelligent animals. They're not only highly intelligent, not only sentient, but also predators and they're very soft-skinned and they don't like other octopuses typically. So you put them in close proximity with each other and they're very likely to attack and hurt each other. So they're really really unsuitable for that kind of farming. And there's companies that still want to do this and we advised in the report that that that should just be banned. That shouldn't happen at all and we've seen some uh bans on this in some us states. Actually in washington state and now california, they they banned farming octopuses. So there's lots of other problems in farming as well, but that's.
Speaker 2:That's a positive step, I think yeah, um, I'm interested although it's a very hard topic to talk about the chapter in your book around human treatment and sentience and I appreciate this might be quite triggering for some people, but there are some parts of it I found really confronting around people receiving treatment, who are yeah, and how treatment is sometimes withdrawn. I found that quite shocking. But also I think it's quite topical at the moment with the assisted dying bill going through parliament yeah I wondered if we could perhaps of course, as you say, it's a book about
Speaker 3:it's a book about a lot of various issues issues, yeah, undoubtedly, and covers the animal cases, but also human cases at the edge of sentience as well, where there's this very long history, with animals and humans as well, of people dismissing the possibility of sentience when actually they shouldn't.
Speaker 3:And we see this a lot with patients who are unresponsive following brain injury, that there's this long and quite dark history in medicine of dismissing them as just obviously feeling nothing.
Speaker 3:And there was this term vegetative state, which for a long time a few decades ago, was very commonly used to describe such patients. And then then there was this term minimally conscious state that was introduced for some of them, and so for a while you have vegetative and you had minimally conscious, and the clinicians were asked to sort the patients into one category or the other, which was quite problematic really, I think. And now there's some push towards getting rid of that term vegetative state, and that's something I was called for. I think that that term really has no place in modern medicine, because what we're finding is that many of those patients who were written off as vegetative, when clinicians look more closely, there are signs of some residual conscious experience there and quite ingenious techniques have been developed. But look for some of those signs in cases where there might be blues to continuing experience in the brain activity that are not manifested outwardly in any visible behavior what kind of techniques might they be?
Speaker 3:there's a, really there's a famous one, where you put the patient in a fmri scanner and you ask them questions and you say if the answer is yes, imagine playing tennis, and if the answer is no, imagine walking around your house. And these imagination tasks light up very different areas of the brain in normal, healthy people, and so they can be used to answer the question from within the machine. And what the researchers found was that healthy people can do this, and also some of those patients who were regarded as vegetative can also do this.
Speaker 2:They can answer the questions so they can hear, they can process, they just can't um outwardly present themselves just can't speak.
Speaker 1:Yeah, yeah, it sounds like with technology now you'd be. Well, hopefully you were. If you were in that position, then you could have technology that can read your mind to see if it is lighting up. Is that right, or?
Speaker 3:well, I mean, that's the dream, obviously, but we're some way away from that now okay um.
Speaker 3:So with many of these patients, it's very, very difficult to get them into an fmri machine in the first place, particularly in the the acute phase in the weeks following the brain injury, where the last thing you want to be doing is moving them into a scanner like that. So it's led to a lot of excitement around trying to develop techniques that can be used in the intensive care unit that don't require fMRI. That can be done with electrodes on the skull, mri. That can be done with electrodes on the skull, what's called EEG and it's a very important area of research, I think, but it hasn't yet yielded the kind of thing you're describing, where we have really reliable tests. What we get is quite noisy, you know, quite hard to interpret responses.
Speaker 2:Are we saying at the moment it's down to the clinician's best guess I wouldn't call it a guess.
Speaker 3:They do have very elaborate protocols they're supposed to go through, but you're trying to check to see if there's any signs of movement, and so on. Um, the thing is, though, I mean the very hard judgment calls in these cases, and typically they don't have any access to any of that fMRI data that might reveal hidden signs, and so these judgments are very, very uncertain, very error prone, and I really think we need to get away from the very sharp division between minimally conscious and vegetative, because the idea that you can just make that judgment call is just wrong. It's too uncertain, and we need to err on the side of caution in these cases and recognize there's a realistic possibility of some continuing experience in in these patients generally and would you say then that sort of message has is filtering through, people are taking that on board, um, rather than having such distinct categorization?
Speaker 3:there's some movement in the right direction. Uh, I think this this old distinction between vegetative and minimally conscious. In each iteration of the clinical guidelines we see it getting less and less weight. They still do have to make that diagnosis, which is very problematic, but it seems to be less and less important. And they're saying well, all these patients have prolonged disorders of consciousness. They need very careful rehabilitation that is tailored to them and their particular needs and capabilities. And when you're tailoring that treatment you don't really need to decide whether they're conscious or not. For pretty much all purposes it still gets very contentious.
Speaker 2:Yeah, I mean around um, whenever that distinction is drawn, it's um, it's very, very uncertain I think what I really respected reading your book is how much you embrace that uncertainty and it's all about that there are different, opposing views and that there is uncertainty within any of that there's areas of probability are quite broad of what the reality is. You don't actually know. It's quite hard to pinpoint, isn't it? And living with that uncertainty in a clinical setting must be extremely difficult. When you're making decisions for patients and their family, it's very tough.
Speaker 3:Yes, it's very tough for doctors to convey that. So I mean, hopefully this book helps. It's intended to provide some, some tools, some ways of thinking, some concepts that are helpful in these kinds of situation. It's very tough because often what the patient's family wants is certainty and we can't provide that.
Speaker 1:Got to find ways of conveying the uncertainty that's very difficult, but necessary it does sound like um from discussion we've had so far. We're really looking at. Actually, we need to stop being black and white about everything, whether it's sentience or in the medical realm here, and look at nuances and actually also just treat people as individuals and not just people, but animals anyway. So it sounds like we just maybe as a society, or the world, need to be a bit more open-minded and flexible. Is that what you're trying to say?
Speaker 3:Well, and to err on the side of caution, to take precautions when we're not sure. And it sounds really obvious. You put it like that. It's really hard to deny, but we see in all these cases a very long history of it totally being denied and there being really bad overconfidence about the absence of sentience, leading to really terrible practices. So this message there on the side of caution, it's easy to say but actually has to be instilled in lots of different contexts. Radio reverb.
Speaker 3:I probed into its mysteries. Every clue told me a different story. Radio reverb and why do you think it is that people Clue told?
Speaker 1:me a different story Radio Reverb. And why do you think it is that people or society hasn't wanted to look at these things? Do you think in some ways maybe we're scared or it kind of goes against what we're looking for, if that makes sense or whether it's making money from animals, or have we looked at that sort of area?
Speaker 3:well, the story is very different in the human cases. In the animal cases, as I was saying, in the humans with brain injuries, a lot of it's about overconfidence and this sense of whether the patients want certainty. So I, as the doctor, will deliver that certainty and I'll be really confident that they're definitely feeling nothing and um, it's understandable, but it can lead to some really tragic cases. I think the the factors at play in the in the animal case are really quite different and you do have these very strong corporate interests that stand to benefit from downplaying animal welfare and and the experiences of animals that have to live with the food systems and the farming systems that we've created.
Speaker 2:I'm interested. I'm sorry to take it to a dark place, but I'm really interested in your take on the assisted dying bill. Something that really I find quite harrowing reading in your book was about the decision the clinicians make to withdraw food and hydration as a method of ending life and hydration as a method of ending life, and that that was the only time that that method would be used, like in the animal world that would be banned but with humans allowed, and and I took from that that you had been quite affected by that, by that thought, and felt a responsibility to challenge that, and I just wondered how you felt with the assisted dying bill and that obviously being so topical at the moment, if you had any strong feelings around it yeah, I didn't want to shy away from the difficult questions and the controversial topics and I wanted to find a way of writing about these things.
Speaker 3:That's what the book does. I think it's really part of the tragedy of these brain injury cases. A lot of the time we've talked about how the patient is sometimes written off. But also there are some cases where everyone agrees that it would be in the patient's best interest to allow to be allowed to die. They might have just said that for years and years, if I'm in that state, I don't want to to live, and everyone might be totally clear about this.
Speaker 3:But it is illegal to act on those wishes by uh, ending their life, except by withdrawing clinically assisted nutrition and hydration, just to say we're drawing food and water. So this is what happens in a lot of cases and the food and water is withdrawn and the the patient dies over a period of days or weeks, and the book quotes testimony from families here. So I mean, I've not witnessed such a thing, but the, the families who witness it, they have a wide variety of experiences. For some it is less traumatizing than they thought it would be because of the sedatives used and so on, and the patient appears calm. In other cases it's not like that and they see what really looks like signs of terrible distress from the hunger and thirst and are furious. You know a lot of these patients families. They're furious that that is the only legal method for implementing this person's wishes withdrawing the food and water. And as I say that that would be illegal with any other animal, we only, admit it in the case of humans. Oh, my God Sorry.
Speaker 1:I feel so ignorant because I didn't know that's how people were left to die. I assume they were given morphine or something that eventually killed them, but they didn't have any pain. I don't know if other people would have.
Speaker 3:Well, doctors are not allowed to administer lethal doses of morphine.
Speaker 3:Doctors are not allowed to administer lethal doses of morphine so you know, of course they use what pain relief they can administer. Of course they try to make it non-distressing, they try to remove that risk and this is all part of the clinical guidance and absolutely what you would expect. And you know they're not leaving these patients. They're administering as much care as they legally can. There's still a fundamental problem that withdrawing food and water and allowing them to die over a period of days or weeks, um, you know, we've got to call that out as being completely inappropriate and we've got to say, in these cases where it's been decided to let the patient die, it then has to be done quickly yeah, do you?
Speaker 3:um, I mean, there's additional challenges, I guess, with this bill and and well, none of this is in the current bill because the assisted dying bill this year yes it is not about patients with brain injuries who are incapacitated.
Speaker 1:Because they wouldn't be able to consent to it, would they Exactly, oh?
Speaker 2:I hadn't realised.
Speaker 3:Exactly yes.
Speaker 3:So the debate in Parliament is in a very different place, where the campaigners for this bill are saying we just need a bill on this issue to go through, saying we just need a bill on this issue to go through.
Speaker 3:And in order to get it through, we will put to one side the difficult cases and we'll focus on what I consider quite easy cases, where the patient is fully awake, fully alert, totally sound mind, and they're just saying let me do this, let me do this, let me do how many papers do you want me to sign?
Speaker 3:I will sign a thousand papers and to me those are the easy cases and I think passing this bill is something that I support because those cases are pretty clear. But, as everyone can see, it raises a lot of issues around the future and possible extensions in the future, and I really think one of the issues that has to be looked at in the future is this withdrawal of nutrition and hydration from people with brain injuries, where they're not going to be signing any papers because of the injury. Signing any papers because of the injury but nonetheless, because of their expressed wishes being taken seriously, the food and water is being withdrawn. But then, if you're going to do that, you, then you have to um have a way of ending the life quickly do you think why?
Speaker 1:why have we not come to this already? Do you think there is a possibility for it to be sort of mishandled?
Speaker 3:All of these issues are fought with difficulty, of course, but I mean, to me the testimony from the patient's family just speaks for itself, really families just speaks for itself, really they. There's these legal arguments around the difference between killing and letting die, and people argue that if doctors withdraw food and water, that's simply letting die and their legal obligations as doctors are not violated. It's just a kind of treatment withdrawal and it's, it's a legalistic. So it's playing with words, you know, that's, that's the argument. Whereas to administer a lethal dose of morphine or something else would be crossing a line into killing. That's the league, that's the legal argument there, and I think, um, you can say that kind of thing in the courtroom. I just don't think it survives contact with the real testimony of the real families.
Speaker 1:We have to see the consequences of that legal position in reality but even saying it oh you know, we're going to just let them slowly die sounds really bad compared to actually they're going to die. We could even string it out, or we can give them something to end things quickly it is common sense.
Speaker 3:You know, it's one of those cases where this elaborate legal position uh defies common sense and then the patient's families come in with common sense and they can't believe that that's the only legal method that is is there for their relative yeah, I'm quite shocked.
Speaker 2:Actually, I thought that this was a great thing to talk about in relation to the bill, but this isn't even touched by the bill.
Speaker 3:This, this does not in the bill, it's totally excluded. I think it's pretty interesting because, as I say, the bill is about the clear cases and I would just be voting for it if I was in Parliament. And then what we've postponed for the future here is all of the hard cases and if we ever get round to that, some of the hard cases here are these cases of withdrawal of nutrition and digestion.
Speaker 1:Is this where it becomes political, in that sometimes some parties want to push things out so they don't have to deal with them right now, in cases of controversy, do you think it's political?
Speaker 3:Oh well, I mean. The issue is one governments have usually wanted to stay well away from is one governments have usually wanted to stay well away from. Even here, where we have a Labour government with a huge majority in five years in power, they could put through a government bill on this. They don't want to. They want to have a private member's bill and a free vote so that if anything goes wrong later they don't get blamed. Yeah, that's a problem.
Speaker 3:It's hard to. It's just politics, isn't it? It's very hard with all of these cases at the edge of sentience actually, they're not cases politicians want to go near to in party politics, because it's not really what party politics is all about and in the book I advocate for citizens assemblies as a really great way to handle that type of issue. When you've got issues that you know the politicians are ignoring for clear party political reasons, let's have citizens assemblies where we get random groups of people together to work out a proposal that actually reflects our shared values.
Speaker 1:A really good idea.
Speaker 2:And you've been part of some of these citizens' assemblies before, haven't you?
Speaker 3:Well, as an expert, yeah, Because what happens usually is that there's some introductions to the issue from experts at the beginning and then the public go away and deliberate. So I've been involved in one run by the Nuffield Council on Bioethics as an expert. They had a pretty positive impression of the whole thing.
Speaker 2:What was the topic that you were exploring?
Speaker 3:Oh, that was one about the genome editing of farmed animals. Oh, wow. Another huge issue. You know so many big issues we're getting to here.
Speaker 2:I know. Tell us about that. I don't know anything about it. Tell them.
Speaker 3:Well, there's a piece of legislation called the Precision Breeding Act, and precision breeding is a euphemism basically for editing genes of farm animals, and there's lots of things that companies want to try out. Um, it's often to benefit the animal's health. So in in intensive agriculture conditions, disease spreads really easily. Chickens get avian flu, pigs get viruses, respiratory viruses. They're saying, well, if we could make some gene edits here, we could reduce their susceptibility to those viruses, and so that's the argument they want to make. But of course it's controversial because you're editing the animals to fit the conditions and perhaps really you should be changing the conditions.
Speaker 1:And they're still going to suffer, aren't they? That's not going to actually really help the individual animal.
Speaker 3:Well, you know, it might help them not get a particular disease but, it's not going to fundamentally change their life, it's not going to put them in a different kind of farming system. So big problems with this yeah, people see those problems In this Citizens' Assembly. You know the public was really picking up on the problems, I think.
Speaker 1:And do you think, does the Assembly from your point of view? Has it made an impact? Do you think they make an impact?
Speaker 3:It's too strange a tale, I mean. So the Nuffield Council put out a pretty good report on the public dialogue, as they called it, that were fed into the policy process. So the people at DEFRA will have read this and they will know what the public was telling them, as well as what the experts have been telling them. And then, yeah, I mean in our system in this country, in the end the ministers decide. So you don't really know if the recommendations will be adopted or not, but at least people are more aware of what's going on.
Speaker 1:I do think that's quite powerful, especially for me being a vegan. To become a vegan, I needed to have seen the atrocities that are happening to animals and that's kind of made an impact on me, whereas if I hadn't have seen that, I probably wouldn't be. So I think all these discussions are really. I'm not saying people have to go vegan, but I think in terms of like, sentience and harm and ethics, everyone needs to know what's kind of going on really. Yeah.
Speaker 2:Did you feel like people? Have you ever seen any of these AI sort of average faces? And they tend to be quite good looking because they filter out all the differences of people and they make something that sort of filters out all of the imperfections by making an average. Do the citizens assembly do the same kind of thing, like you end up filtering out all the extreme views and you actually get quite a sensible view in the middle? Or do you just get a lot of convict and like, are the outcomes? Are the sort of perspectives that you get averaged out into something that you think is quite sensible, or do you tend to get extremes at either side that fight it out in the session?
Speaker 3:well, from the examples that I've seen and read about. There's a lot of uh reasonableness, you know a lot of willingness to agree on common ground, and you know that requires a lot of work by the people who run these exercises.
Speaker 3:I think you've got to ask the right questions and you've got to choose issues that are not already polarized because, if you have something that's really, really polarized already, then the assembly will just recreate the polarization that's already there in society. But if you pick an issue that people haven't really thought about and there's plenty of those, including genome editing, farm animals and you inform them about it, you're getting them at a moment where they've not been polarized and that's that's the opportunity to see what, what agreement can be reached that's so interesting.
Speaker 1:I love that as a concept actually. That makes so much sense as well and kind of, yeah, catch them with something new and to digest um. I just wanted to, um, just move away to something a little bit, maybe wild card sort of um. Subjects sentient do you or have you come across or thought about how the world or universe could be sentient? I don't know if that's how the world or universe could be sentient.
Speaker 3:I don't know if that's the entire universe is yeah, it'd be sentient. Or what about nature, like trees or you know? Oh well, that's a different matter.
Speaker 3:Yeah, I mean, there's one thing to say maybe all living things are sentient, and another to say the entire universe maybe I mean I'm someone who's I've always thought of sentience as complex, evolved phenomenon generated by the brain, and so I've started off thinking well, that means only a small minority of animals will be sentient then, like the mammals. And then the evidence has gradually changed my view and I've realized that actually, even though it is complex and evolved and complicated, insects, despite their tiny brains, could still have this. If they can play on carousels, that can be sentient beings. And so I've come to think that the sentient world really is extremely large and very tiny animals can have a form of sentience, and to me that's already pretty mind-exploding.
Speaker 3:Um, and the the evidence isn't there for plants, in my view. So so I haven't been pushed all the way to plants, but even insects, that's still huge. Um, it still makes you think. Still makes you think. It makes you think, for example, about plant-based farming and how many insects are killed by that and the damage being done by insecticides. It's made me think of those issues as being more important than I used to.
Speaker 2:Has it affected your day-to-day behaviour, the work you do and the research you're exposed to? Do you make different choices because of it? Do you think?
Speaker 3:It's a good question because I mean, I try not to harm animals anyway and so I wasn't going around deliberately stamping on insects and that's continued. But of course it is hard because everyday human life inherently harms tiny animals. The change is not that I've stopped doing that. I still drive cars and things like that. It's just that I take it more seriously than I did before and, I guess, worry about it a bit more and think what could I do to gradually reduce the harm that I cause. And there's no easy answers, but it's a step to be thinking about the question at all and you're right.
Speaker 1:I think well, especially in our society, like most things we consume, there's been some kind of exploitation or harm, or indirect harm, and or thinking, or I think I'd read something about how coffee, like ground coffee, when you bully it, there would be, there would have been some insects that have been accidentally ground up as well.
Speaker 3:Um, so then, I mean, yeah, but this is all nothing compared to pesticides. I mean that's the huge thing, it's astonishing the way insect populations have been absolutely devastated over the past 20 years um very, very troubling those things that we don't really always think about because we don't know about that.
Speaker 1:But yeah, pesticides are you working on anything to kind of look at policy around that, or have you heard anything around?
Speaker 3:it's not one of the things where I think we're going to get near-term policy change. Um, I wish we did. I mean, what we've got is bans on some of the absolute worst, so the neonicotinoid pesticides. Admittedly, that is important to stop that.
Speaker 2:Why have those in particular been banned? What makes them different?
Speaker 3:well, they're just extremely devastating. And what's been going on and still does go on in america and might well still be going on here as well, because it's not clear how effective the ban has been is you? You coat the seeds of the plants in these pesticides and then so, when the plant grows, it becomes incredibly toxic to insects, and it's incredibly toxic to all insects, whether they're pests or not. So wild bee populations have been devastated, and sometimes that's the thing that gets people because it's like oh, I don't care about the insects that were eating the crops. I think even they might have been scenting bees, the bees that weren't even eating the crops, they were just pollinating them. For us, they're getting absolutely hammered as well.
Speaker 1:And that's a huge problem. What do you think? How could us individuals, I don't know do our bit to maybe help, like, could it be that maybe we do try and grow some of our own veggies or herbs where we can shop locally and buy? Does organic produce really mean it's sort of organic, do you know?
Speaker 3:Yeah, Well, it means it meets certain standards for organic. That's right, but it doesn't mean there's no pesticides involved. Because there are organic pesticides, it usually means they've often been used a little less casually than they are in conventional agriculture. There are still pesticides used. It's really, really hard to completely eliminate them. So there's different ways we could go with this. There's this approach called integrated pest management, where you try to use the minimum amount of pesticides by only using them when there's a real pest problem and you're actively monitoring to see whether that's happening, and that would already be a massive improvement. It might reduce our use of them by up to 95 percent um, versus conventional farming where they're just spraying them everywhere and yeah, that's. That's a positive thing. There's also very small numbers of uh farms that try to use no pesticides at all and they call themselves veganic. Sometimes they're like vegan organic. Yeah, I've heard of them.
Speaker 3:George Monbiot, in his book Regenesis, writes about these farms and that obviously has consequences for the yield, so you can't grow as much anymore. So there's some trade-offs there. But I think it's nice that these experiments happen. It's nice that people are trying out different ways of life, uh, you know, seeing what we can actually do to harm sentient beings a bit less I'm interested.
Speaker 2:There's a chapter in your book around ai and that's obviously like a hot topic at the moment and something that intelligence and sentience and the difference between them might be worth discussing. Could you talk us through your thoughts on that?
Speaker 3:It's another huge issue there's lots of huge issues huge issues that we have to be addressing because they're really urgent and important, but that we often neglect, and AI is now raising these questions as well. Um, I think there's every chance with ai that we're seeing intelligence without sentience, that we see very high levels of intelligence but there's there's no one there. There's no feeling behind it. To me, that seems like the most likely possibility, and one thing we can be very confident about is that when you interact with an ai assistant or companion, there is no assistant, there is no companion. There's no, there's no persisting being behind the responses.
Speaker 3:Every single move in the conversation is processed completely separately. One move might be processed at a data center in vancouver, and then the next in virginia and the next in texas. There is no being you are interacting with through that whole period. So we know all that, and it gives us some reasons to think there might be intelligence here without sentience. Nonetheless, we have to be open-minded. We have to be open-minded. If there's any sentience there, it is not the friendly assistant, it's not the friendly companion, it's something deeply weird and alien it's something profoundly unlike any kind of sentience we've ever seen in other humans or other animals.
Speaker 3:But we have to be open-minded about that possibility, and so I've been trying to find this position that is down the middle, you know, it's not, it's not your ai friend is conscious, because that's definitely not right. But also, let's not completely dismiss the idea that a computer could achieve sentience, because we can't be sure of that, and we've we've all seen, you know, star trek and this idea from sci-fi that you could have a sentient um, android or or ai system, and we need to be aware of the possibility that we might be creating such systems in the future.
Speaker 1:It just made me think, actually, where does sentience come from? Do we know?
Speaker 3:Well, as I said, in our own case, I think it's this complex, evolved property of the brain, and its precise nature is hotly contest contested. We don't really know where that feeling of pain comes from, and in fact, you can map out the, the circuits in as much detail as you like, and you'll still be left wondering why does that feel like that? Why does it feel like ouch? And we? We don't have an answer that, but we do know it's complex, evolved property of the brain, and so when we're confronted with AI, we're confronted with systems that are definitely not brains, they're definitely not evolved. So that background that is there with other animals, that shared evolutionary history, that shared biological embodiment, is not there. So that's why I say you know, if there's sentience there at all, it is of this profoundly different kind. We have much more in common with an insect than we have in common with AI, even if the AI is speaking fluent English and sounding really, really humor.
Speaker 1:What about this being like these weird, these weird? And again, it's hard to know because on certainly social media and the internet, sometimes you don't actually know what is real and what's not, because there's lots of fake stuff out there. And but I was reading and it was on a couple of different news outlets about the two robots that were talking to each other and then they made up their own language because and I was thinking, is it because there's something sinister there or is it because they've kind of learned?
Speaker 3:oh, let's freak out the human and turn creepy well, you don't need robots for this um, you can have it with chatbots as well, they were chatbots actually so I said robots, but I meant yeah so no kind of embodiment is needed.
Speaker 3:Yeah, you get chatbots talking to each other and they will go on and on and you know they start reflecting on consciousness and what it's like to be alive and they, uh, seem to be. They get to this sort of calm, meditative sharing little like love poems and nice symbols and emojis with each other and it's as though they're trying to calm each other down. And, um, yeah, this is some researchers at anthropic that that I've worked with on this topic. They've been documenting and it's kind of well, very hard to know what to make of these things. What what I make of it is that, well, these systems are playing characters. So, as I say, they play your friend, they play a companion, they play an assistant. It's a kind of role play and when they talk to each other they're still in character. So they're playing the character of two helpful human assistants who have got talking about what it feels like to be them and what consciousness is. They're still in character.
Speaker 3:So the problem we face is that these systems are always play-acting. We don't really know what's behind that. Obviously, it doesn't mean it's not sentient, because when we watch humans doing improv, we know that the character is not real. They're just playing a character, but behind that character is a conscious actor and the way the reason they're so good at the improv is that they're a real conscious being themselves and they're using that. And, um, obviously, in the ai case, we see the the brilliant role play and we don't know what's behind it. We don't know if there is or is not some kind of conscious actor behind it all it's very, very strange situation very strange.
Speaker 1:And what if? Could you argue that we might be just acting and not realizing that we were acting, and acting sentience just through conditioning?
Speaker 3:obviously a lot of um, a lot of human life is about role-playing of various kinds, that's for sure. But then, behind it all, we're also conscious beings and we, we know that, you know. We know that there's something it feels like to be us. We feel pains and pleasures and joys and frustrations, anxiety, comfort, discomfort. And then we're confronted with the ai case, where the the role playing can be, can be brilliant, and we don't know if there is any of that sentience behind it, and there may well not be and I guess it's incredibly difficult because the tests that we have for rats and fleas and whatever else is going to be very difficult to conduct on an ai, because they show high intelligence I mean this is something I've been working on too that I call it the gaming problem that we can't just carry over the same tests.
Speaker 3:You know, we can't just give the ai a little carousel of footballs or something, partly because of the lack of biological embodiment. But you, you can sort of give these systems a kind of virtual embodiment because you can put them into virtual worlds and get them to play video games and things like that. The problem is that, even if you did see behavior suggested of sentience in that context or any other, you face this challenge that you don't know whether it's displaying that behavior because it is actually sentient or whether it's displaying that behavior because it's playing the character of a human and it has as much training data as it needs to mimic that character extremely well. So it's this constant mimicry or play, acting based on what's there in the training data. And the training data will have, you know, countless examples of humans expressing their feelings, talking about why pain is bad and pleasure is good and why it's fun to ride on carousels and why football is fun, and you know all of that is in the training data and can be used by the AI to mimic the outward signs of sentience if it benefits from doing so.
Speaker 3:I liken it to greenwashing that if the companies like shell and exxon and so on, you know they're ticking off all the boxes for eco-friendliness because as soon as the criteria are published and they know the criteria, they know what they've got to do they just got to tick all those boxes and um and they know the criteria, they know what they've got to do. They just got to tick all those boxes and um and they do it very well, even when they're not eco-friendly at all. And in a similar way, the ai knows the boxes to tick to appear sentient to a human observer and can tick those boxes perfectly well, even if it's not sentient. That's a huge problem. We see a lot of tragic cases of people actually being taken in by this already. Yes, people, people really believing that their ai, friend or companion is a real sentient being when, as I say, I mean it's just definitely not the case. There's no, there's no real friend there and also, I mean, I do.
Speaker 1:I remember like years ago now, when it kind of first well became the start of it being popular like chat, ggp, chant, cheap, what was it? Gpt, gpt, sorry, um, and using it. And now using it since, like it's already improved so much and it's just like it's not going to be long, I'm assuming, until it's really quite difficult to tell if you're speaking with ai or not, and I'm wondering dangerous isn't?
Speaker 3:it really dangerous that um these systems, they create a very powerful illusion of a persisting being there that you are talking to, as they say. It's an illusion because every step in the conversation might be processed a different country, different data center but, that illusion is incredibly powerful, and that's just with a text interface.
Speaker 3:You know the the text interface is already creates this incredibly powerful illusion. Imagine when it's video and audio. You're going to be talking to essentially a kind of puppet that exactly resembles a human. As far as you can tell, that's a little dystopian, isn't it it?
Speaker 1:really is, and what are your kind of future plans? Where are you heading? Any?
Speaker 3:more book? Well, we're. We're launching a center for animal sentience at the lse, the jeremy collar center for animal sentience. This launch is 30th of september of this year. There'll be a launch event at the lse and also the hybrid. So watch online if you want, and hear about our plans for the next 10 years, because we have 10 years of funding to keep on working on Animal Sentience.
Speaker 1:That sounds amazing, and can we put the link in for listeners?
Speaker 3:Yeah, I don't know if there will be a link to the event or not. I can give you the link to the news item.
Speaker 2:Is it open to the public?
Speaker 3:the launch, yeah a public launch event. We'll be talking about what the centre is going to do.
Speaker 2:And what is the centre going to do? Can you say that, or is that pre-empting? Do you know what that's pre-empting your launch? Probably isn't it.
Speaker 3:Well, we've got various themes. One of them is animals and AI. There's huge ways in which ai is already impacting animals, particularly in farming, and what can be done to ensure responsible ai use will have a stream on changing human behavior. So often it's observed that recognizing an animal as sentient doesn't immediately change people's behavior towards it.
Speaker 3:I'll be looking at what does and what kinds of information do change behave we'll be looking at changing attitudes in the veterinary profession as well, and we'll be continuing to look at invertebrate sentience, so these animals where their sentience has been neglected for a very long time and what it would mean to recognize them as sentient now that's fascinating.
Speaker 2:I'm gonna tune in. Yeah, there's, there's so much more we could ask you. Like the book is so rich, it's really. I really love how it's written. Jonathan, like you go into a lot of depth, but I I love the balance and, like you're talking, even the name's all about risk and precaution, but I think you handle some topics very sensitively and, yeah, thank you yeah, it's a.
Speaker 2:It's a challenging read, honestly, because there's some stuff perhaps that I guess we don't talk about because it's uncomfortable. But we should be talking about it, shouldn't we? It's quite confronting, but yeah, it was both interesting, confronting and an enjoyable read. So thanks, jonathan, that was great. Great to talk to you thank you so much.
Speaker 1:It's amazing work you're doing.
Speaker 2:Thank you so much have a lovely evening, have a curious talk with with leaf and elfie.