>> STEPHEN KUEBLER: Hello and welcome to "Ethically Speaking," a UCF interdisciplinary series on contemporary moral issues. My name is Steve Kuebler, and I'm the organizer of this series. I'm professor of chemistry and optics and founding associate director of the UCF Center for Ethics. In this inaugural lecture we'll learn from Professor Michael Levin about how artificial life is becoming a reality and what the ethical implications are for our world. Before we introduce Dr. Levin, I'd like to take a moment to tell you about this speaker series and how you can participate.
I have the pleasure of collaborating with Dr. Jonathan Beever who is associate professor of ethics and digital culture at UCF and founding director of our Center For Ethics. Together we developed "Ethically Speaking" to promote conversations about challenging issues of our time.
We encourage you to visit the center's website at ethiccenter.research.ucf.edu to learn more about our activities and to find out about other speakers in this series. We'd also like to hear your feedback and your ideas for future talks in "Ethically Speaking."
Ethics is important to us at UCF, and that's clear from the long list of partners who co-sponsor this series. All of us in this partnership hope you enjoy the discussions in "Ethically Speaking," and we hope it cultivates understanding of how ethics provides frameworks for making better decisions, particularly when there are competing values. Today's presentation is being recorded and links will be posted through the Center For Ethics so you can share it with others.
At the end of his presentation Dr. Levin will take questions, moderated by Dr. Beever. You can submit questions using the Q&A box, and you can also up-vote questions that interest you. Now we will turn to Dr. Will Crampton, UCF associate professor of biology, who will introduce
our speaker. >> WILLIAM CRAMPTON: Thank you, Stephen, and good afternoon everybody. It is a great pleasure to introduce today's speaker, Dr. Michael Levin.
It's also an honor for the biology department to co-sponsor this important talk series. Dr. Levin is distinguished professor and Vannevar Bush professor of biology at Tufts University and also director of the Paul Allen Discovery Center and the Tufts Center for Regenerative and Developmental Biology. Dr. Levin obtained dual bachelor's degrees in biology and computer science at Tufts University. He then completed a doctoral degree in genetics at Harvard University as well as subsequent postdoctoral research in molecular embryology at the same institution. Dr. Levin then held several prestigious academic positions including an associate professorship
at Harvard before occupying his current position at Tufts in 2008. Dr. Levin's research career has been enormously productive. His publications have spanned the fields of developmental biology, biophysics and genetics and have been cited almost 20,000 times with an H factor of 75 and an astonishing I-10 index of 256. His work on the molecular genetic mechanisms that allow embryos to form left/right asymmetric body structures is on the journal Nature's list of 100 milestones of developmental biology of the 20th century. One of the most exciting recent developments from Dr. Levin's research program, which we'll
learn more about today, is the creation of a kind of artificial life in the form of microscopic programmable virtual creatures called xenobots. Dr. Levin, the UCF community is delighted you could be with us today. The floor is yours. >> MICHAEL LEVIN: Thank you very much for that introduction, and thank you for the opportunity to share these ideas with you all. I am not a professional ethicist or philosopher, so my goal today is not to provide any answers or make any specific ethical claims, but what I do want to do is to go over some recent advances in bioengineering and regenerative medicine which I think have some very important implications for working out ethical principles going forward. If anyone is interested in contacting me afterwards or you want to see the details T deep biology this is all based on, you can find me.
I would like to make explicit the lens with which I view the world because I think it's important to see these things be made clear. I think that science is the best window we have on the world, and I think that what ought to be should be a very tightly coupled to our very possible picture of what is. I think philosophy has to keep up with the science. I also take a kind of engineering approach in my group where we study the natural world, but we believe that a mature understanding must encompass the space of the possible, not only whatever happens to be true right here and now, but actually all of the adjacent possible facts that could be made to be true as science moves forward.
So the main points that I would like to transmit today and hopefully discuss with you all are these. The first is that I would like new technologies to be evaluated from the perspective of opportunity cost of failing to improve life for all. I will try to point out that evolution offers no support for the common assumption that our world is basically great and all we need to do is not to make it worse.
This is a common thing many people in ethics discussions of new biology ask, isn't it scary? What happens if this and that bad scenario takes place. I think this absolutely has to be balanced with the idea that, no, we actually have an opportunity and a duty, a moral duty to improve the human condition with research. This is a transition to an opportunity-based view as opposed to a purely risk-based view. I'll talk about what it might mean for things to be natural or unnatural for ethics. I'd like to view as how bioengineering as well as taking evolution seriously is going to demand a revision to very basic concepts that support theories of ethics. All interesting qualities, intelligence, agency, responsibility and so on.
In fact, lie on a continuum, and we cannot ask which types of systems, these include living agents and new synthetic creatures are or are not intelligent, are or are not cognitive in various ways, but we need to ask what kind of cognition they have and how much. There are no binary categories supported by science. I'll make the claim that we're out growing this very binary terminology of robot, machine, cognitive agent. These binary terms are not serving us very well.
We need to move beyond them. What I'd like to do in the first part is show a number of technologies that people have described as scary. I hope I can show why they're, in fact, essential. So one thing to realize is that pretty much all biomedical needs so everything with the exception of infectious disease, things like birth defects, traumatic injury, aging, cancer, degenerative disease, would be solved if we knew the answer to one question.
That question is how do we control what cellular collectives decide to build? So groups of living cells get together, they cooperate and build specific structures, complex organs, whole bodies. If we knew how to tell them how to build specific things, healthy organs and so on, we could solve all these biomedical problems. It would be an enormous advance.
The future of our field I like to illustrate with this which we call a anatomical compiler. At some point in the future, what you ought to be able to do is sit down in front of the software and draw at the level of anatomy here in this schematic, not from the level of molecular pathways or those kind of details but draw the plant or animal that you want. In this case we've drawn this three-headed flat worm you'll hear much more about flat worms shortly.
If we had a proper understanding of how cells decide what to build, we would be able to have a piece of software that would take this anatomical description, I want eyes over here, a third head and so on, and compile that into a set of instructions, a set of stimuli that would be given to cells that would build whatever you just drew. Basically computer-aided design for living forms. You should be able to create any living form you want, any organ of the body, any novel creature that's never existed before. Now, I should point out we are nowhere near close to this general capability, although we have a few inroads here and there, but this general capability is still far away. Let's take a look at why it's actually so far at this point. We all start life as a single cell.
This is a fertilized egg. This gives rise to one of these remarkably complex anatomies, quite reliably. In fact, it's really important to understand how this works and it's important to know this is not simply a stem cell biology problem. For example, this is a teratoma, a tumor that may have hair and skin and teeth and muscle and bone.
And what makes this different from one of these wonderfully anatomically normal animals is the three-dimensional structure is missing. The individual stem cells created all the differentiated cells you want. It has all the tissues.
But it's missing that large scale three-dimensional structure. This here is a cross-section through the human torso. You can see the incredible order, the complexity with all the tissues and organs in the right place, the right orientation, relationship with each other, right size and shape. Where does all this pattern come from? Where is it specified? Everybody starts off as a collection of blastomeres, these early embryonic cells.
We can read genome and it says absolutely nothing about this directly. The genome specifies proteins. What's in the genome is specification of the micro level hardware that every cell gets to deploy. The genome doesn't show any of this information directly. So we still need to understand how does the hardware that's specified by the genome end up making these regular collectives that know what to make and when to stop growing. As workers in regenerative medicine, we would like to know, if a piece of this is missing, how do we convince the cells to rebuild? As engineers, we say what's possible to build, can we get the same exact cells to build something completely different.
If so, how would we do that? So, single cells, in fact, are extremely competent. What you're seeing is a single cell animal known as a lack mayor yeah. There's no cell to cell communication. Yet, it handles all of its physiological, behavioral, morphological and its metabolic needs at the level of a single cell, the scale of a single cell. Remarkably when cells join together, they didn't lose this intelligence. They, in fact, scaled it up and deployed it toward higher order goals Here is a tadpole, here are the eyes, the brain, the nostrils, the guts.
They need to become frogs. They have to rearrange their face. The jaw haves to come forward, the eyes have to move. Everything has to move. It was thought that in some way the genome provided a hard-wired set of movements for all these organs.
After all, all tadpoles look the same, all frogs look the same. We made what's called Picasso tadpoles, the eyes are off, the jaws at an angle, eyes on top of the head. Everything is in the wrong place. What we discovered remarkably is these animals still make largely perfectly normal frogs because all of these organs move in novel paths, and sometimes, in fact, they move too far and have to double back to land in the correct orientation. Everything keeps moving around until you get a correct frog.
That's when the system stops. In fact, what the genetics gives you is a machine that's able to do this amazing error minimization scheme. It's able to meet William James' definition of intelligence, being able to get to exactly the same goal by different means, despite starting off in the wrong configuration, it can still get to its goal through novel unexpected actions.
If we had a robotic swarm that was able to do this, we would certainly call it a kind of collective intelligence. We'll come back to that momentarily. That plasticity, that ability to get to the correct structure despite perturbations can be seen in the structure of the nervous system.
If we make a tadpole with no primary eyes, but we put an eye on its tail, these are real live animals. We put an eye on the tail and use this device to ask if that animal can learn to see out of that eye. In fact, they see perfectly well. We can train them on visual queues.
This is the basis of sensory augmentation technology for humans. This brain, which evolved for millions of years to use visual input from regions where the eyes should go, now finds itself in a novel configuration, where the eye is back here, connected to the spinal cord, not the brain. It has no problem using that data for behavior. The plasticity is incredible.
The ability to handle novelty in living tissue is amazing. One of the things that our laboratory is doing is trying to understand how do we- how can we control these - how can we control these decisions that are made by multicellular collectives. In controlling those decisions, how can we produce novel anatomies. This is all part of trying to learn to control anatomy for various purposes.
You can see here a couple of examples. Here are some two-headed flat worms. This is a flat worm that normally only has one head. We actually manipulated the electrical conversations that these cells are having with each other and they built a head with two animals. It has exactly the normal genome, but the electrical conversations between those cells, we have manipulated them to say make two heads instead of one. Similarly, here is a five-legged frog where we used light to change some of the signaling, the electrical signaling property of cells in the frog's face.
Actually we can induce the growth of a whole limb here. You've seen a five-legged frog and some two-headed flat worms and a tadpole with an eye on its tale. About 50% of the phone calls I get suggest these bodies are new, they're different and they're a little disturbing.
People are used to things that are produced very slowly, by evolution over geological time scale. This kind of thing makes people nervous. These are very new anatomies. One of the important things is that we not only have the opportunity to control anatomy, I'm going to claim we have an absolute moral responsibility to do so. So here is some general statistics. 2 million people in the U.S. alone are living with loss of limb, and the mortality in the
next five years after loss of limb can be 80%. 10 million people die from cancer every year. 50 million Americans are affected with degenerative disease.
Every 4 1/2 minutes a baby is back and forth with a birth defect. And a remarkable 77 billion land animals are slaughtered for food every year. So this idea that our job in bioethics is to simply make sure we don't screw things up, this idea that basically everything is fine and we need to be careful that we don't make things worse is really I find extremely insufficient.
These incredible needs, this global suffering accounts for the other 50% of the calls that I get which is just an unbelievable range of human patience that require all sorts of solutions that we don't currently have. So I think we need to focus on the opportunity cost of failing to live up to this moral imperative, to use science to improve life for everybody. A lot of people will always ask - they ask me what keeps me up at night in terms of these wild technologies. What keeps me up at night is the possibility of fail your to live up to our potential, to use the science that we have access to to make life better. So I want to point out that - I wanted to dispel this idea that there's any necessary correlation between what's actually natural, meaning what was the result of the natural evolutionary process on earth and what actually is what we want to have happen.
This is the basic evolutionary process. It's very simple. The strategy is this, generate lots of offspring which resemble but not perfect copies of the parents.
These are random errors due to changes in the genomes, and the offspring will compete for finite resources. Only some will live to reproduce. They do this again. Generate more offspring, more random copies.
Some are better suited than others, and on and on this goes. The important thing to note about this process is what it's guaranteed to optimize is the quantity of biomass that you'll observe in any environment. That's all.
Nothing in this process optimizes for happiness, for well-being, for any of the things that we as moral creatures should care about. All evolution optimizes for is biomass in any given environment. I think this provides a really important lens on this question of what's natural, what's unnatural and what we ought to be doing. There's nothing magic about what comes out of the process of evolution.
Nothing here is optimized for the things we find of value. We should definitely be able to do better than randomness. I want to show you examples of the things that may be within our reach.
Here is what might be possible. This is a Mexican salamander that regenerates its limbs, eyes, jaws, portion of the brain, heart, tail including spinal cord, muscle, ovaries. It basically looks like this. If a limb is amputated or one of them bites off a limb, which they do to each other all the time, this system will continue to grow, and it will grow exactly what's needed and will stop when a correct salamander limb is formed.
This is one of the most critical areas of research right now is to figure out, first of all, how to start the process, but even more so, how does it know when to stop? How does it know that a correct salamander limb has formed? This is the kind of ability that is possible. Evolution happens to have derived an animal like this, a complex vertebrate that does this, so we know this is possible. In fact, here are the plenary. Plenary are the champions of regeneration. Even though it has all the same neuro transmitters that you and I have, they have this amazing ability, if they're cut into pieces, each piece will regenerate exactly what's missing, no more, no less, and give you a perfect little tiny worm. In fact, that's how they reproduce.
They're so regenerative, they're actually immortal. They have no life span limit that anybody has ever seen. They appear to have conquered aging entirely.
So then having seen this and having seen the salamander, one wonders why is the human life span somewhere around eight decades, nine decades. There's nothing magic about that number. It could have been 12 or it could have been 300, or we could have been immortal like these other animals. There's nothing magical about the status quo. It's our responsibility to try to improve it. Humans have some regenerative ability.
Here even the ancient Greeks knew the human liver is highly regenerative. Deer regenerate meters of bone, about a centimeter and a half per day. All the limits on human repair capabilities are purely arbitrary. They're the outcome of all kinds of basically satisfying decisions that evolution has made and kind of the meandering course of evolution's search through the space of possible outcomes in biology, is what we ended up with it.
It doesn't have to be that way. There's absolutely nothing special about where we are. Our duty is to control for better outcomes. I want to show you what this looks like. So this, of course, is the evolutionary process that I started out by telling you individual cells can learn to work together toward large scale goals.
Here is a goal. Make a limb and then stop. That process of scaling up the goal-directed activity of cells can actually break down.
The scope of the goals that these cellular collectives follow can shrink. This is a terrible brain cancer called a Glioblastoma. They are still very interested in single cell goals like proliferate, make many copies of yourself, go where life is good, take up nutrients and so on. They're completely uninterested in the large scale goal of make a healthy liver or brain or limb. It turns out that, if you work on this question of how is it that cells cooperate with each other in the first place, you can do things like this.
This is a tadpole injected with a human gene. Even before the tumor becomes apparent, you can use voltage imaging - this is a way to map out the electrical activity in this tissue so you can see the electrical conversations the cells are having. You can already see in this region these cells are going to disconnect from their neighbors, require a bizarre electrical state and basically uncouple from the network and treat the rest of the animals, external environment. Once you understand this, you can try to gain control of that electrical process. When we do - this is our work from a few years ago.
When you do this, despite the fact that this oncogene, there in fact doesn't need to be a tumor if you manage the bioelectrical state appropriately. You can reconnect the cells and get them to work towards making things like normal skin, normal muscle. So normalizing or preventing tumors, not chemotherapy to try to find the balance between not killing the patients and killing the tumor cells, but actually to normalize them. That's one kind of advance. This is a frog leg.
Frogs, unlike salamanders do not regenerate their leg. Having lost a leg, 45 days later, there's basically nothing. We've been able to come up with a treatment that triggers a bio electric state to tell the cells, build a leg here.
When you do that the leg starts to grow. Just 24 hours of treatment with the bio electric drugs kick starts over a year of leg growth. You can see the leg is both touch sensitive and motile and it grows for a long time. We're now in efforts to transition to mammals, hopefully some day to humans with limb loss.
David Kaplan and myself are co-founders of Morphoceuticals, Inc. That's a disclosure. There are also ways to address brain defects or other types of birth defects. So this is a normal tadpole. This is an animal that either has a mutation in an important gene, like notch, or it's been exposed to alcohol or nicotine or one of many other kinds of teratogens. You can see the forebrain is missing, the mid brain and forebrain is a bubble.
They lay there and do nothing. It's a terrible birth defect. What we can do now is, guided by computer models that we've created, we can actually provide very specific drug treatments with drugs that turn on and off the different electrical activities that the cells are using and convince them to make a brain of the appropriate shape and size. When you do this here, you find you rescue the brain's shape, the gene expression in the brain and the behavior. They get their IQs back to basically indistinguishable from normal. We have the opportunity to use these technology and perfect them in these model systems.
Ultimately the goal is to use all of this for bio medicine, both in human and veterinary use to relieve the kind of suffering that is out there from all these cases. So that was part one. Part two of the talk, what I would like to stress is a couple of interesting concepts that are really brought on by advances in this technology. The first one is the idea there's a continuum of agency. That properties like intelligence and moral responsibility and memory, all these kinds of things, exist on a continuum.
They're not binary quantities. I want to talk a little bit about the plasticity of individuals during their lifetime. Then I want to show you some really novel beings that have never existed on earth before and we can talk about what that means for us. The first thing is, when we talk about cognition, we don't only mean a full-blown, human-like self-consciousness. That's somewhere up here at the end.
But all the way down to very simple creatures, you see different kinds of cognitive capacities, so predictive or non-predictive, with feedback or without feedback, purposeful and non-purposeful, active and massive. This is a very smooth kind of continuum where there are certain capabilities that emerge, but all along here there are different ways to be a cognitive creature in the world. I want to show you a couple of examples. This is a slime mold. This slime mold - slime molds are uni cellular.
Even though it's huge, it's one cell. What you'll see is it's going to solve the maze. Here are oats over here, more oats over here.
It wants the shortest possible path to the oats so it doesn't waste energy elsewhere. As soon as it's found the right path, it basically withdraws from all the other parts of the maze that are not part of the solution. Here it is.
You can see here it's pulling away from regions it knows are not the right part of the maze, and it does this by virtue - you can see here the flows, these fascinating signals that flow through the structure of the slime mold. You can see here it's a perfect solution. This thing has solved the maze, found the exit, from the entrance to the exit. It has the capacity for certain kinds of problem solving. This is our work from a couple months ago.
Here it is. It learns about its environment by tugging on the substrate, and it can feel there are three glass disks here instead of one. It already in the first few hours of growth, it already has an internal map of its environment, and it will always outgrow towards the larger mass.
So it has the ability to probe its environment and make these simple kinds of decisions. What's really critical to understand and what modern biology is telling us is that, if we take evolution seriously, there is no bright line separating any kind of - any kind of distinction between the cognitive capacities of creatures. So people often ask a fish or something like this, is it conscious or does it have memories, does it have goals? This is a bad way of framing the question because it's not a binary category.
There are no binary categories here. We are all part of a continuous ascent along this kind of structure, and all of these things should be asked in terms of what kind and how much. So what kind of cognition, how much cognition, what is the internal world of any given creature. There's no place to draw a bright line.
It's a little bit like the notion of being an adult. There's a convenient fiction that society follows where, on your 18th birthday, bam, you cross over this line. We all know nothing happens on that day.
It's a continuous growth. This is actually true for all the things we're interested in. There are no sharp lines. We know this even - we've known this for a really long time. Let's take a concept such as personal responsibility or something like this.
If you feel modern humans have it, all you have to do is walk backwards and ask yourself at what point - if you think it's a binary kind of quantity, at what point did you have an offspring that had it and parents that didn't, right? That's what you would be committed to if you thought there was some bright line it might be something that develops quickly compared to the geological time scale, but in terms of the biological, in terms of having parents that don't have some sort of cognitive capacity and offsprings that now do, biologically it's not viable. So we have to understand all of these things are slowly and gradually changing from more primitive versions to the more complex. People talk about sometimes the collective intelligence of ants, ant colonies. It's important to recognize we're all collective intelligences.
We're all made of cells. No creature can be a single, indivisible cognitive agent. We're all made of parts. Here is a single cell chasing after a bacterium.
I'll play it again. This is a famous old movie of a single cell. This is a human immune cell. Our bodies are made of things like this. Our brains, in fact, are a collection of neurons, of other cell types, specialized, but still individual cells.
And we are, and whatever moral, ethical, personal perspective you have looking outwards, you have memories, goals, a personality. All of this is emergent from the collective activity of the cells that make you up. In fact, those cells are - the tissues and organs in your body are, in fact, composed of cells which themselves are composed of sub cellular components that have their own dynamics of those of molecular components and so on.
So there's this massive multiscale type of architecture that biology uses where there are active processes at every level of organization and every scale that have their own local goal-directed activity and all of this somehow percolates upwards in ways that we still don't understand. Because we are all made of parts, this is why change does not have to be taking place on evolutionary time scales. It can take place during the lifetime of an individual.
Here is an example we've known for a long time. Here is a Caterpillar. The Caterpillar has a brain that is perfectly suited to driving this soft bodied creature.
There are no bones. So driving this kind of body is particularly interesting. It crawls and it chews plants. Then this has to transplant into a completely different hard body creature that will fly and drink nectar. During metamorphosis, most of the brain is basically liquefied.
A lot of the neurons are broken down, the synaptic conditions are broken and a brand new brain is rebuilt. So one can ask really interesting philosophical questions. Is it the same creature? If your view of personal identity is the continuity of body structure, sort of that's the case.
Some of the cells are the same. Some of the materials are the same, but many are not. But if your view of personal identity is as contiguous memories, then this conserves identity because it turns out that the butterfly still remembers things that the caterpillar has learned, even though the brain is completely remodelled during this. You can learn more about that here. The brain of an agent can change radically, and one can wonder what is it like to be an agent. You can ask what's it like to be a caterpillar or butter flew, but more interestingly, what's it like to be a caterpillar changing into a butterfly.
In planarian, you can train them to where they get fed and then you amputate the whole head. The tail sits for a week. It grows back a new brain and new head and still remembers the information. That information is transferred from the rest of the body - I'm printed on the new brain. If you think this is something weird that planarian and Caterpillars do, keep in mind lots of people are developing stem cell therapies for human degenerative disease in the brain.
So probably at some point soon we'll have human patients with 50, 60 years of memories, having chunks of their brain replaced by naive progeny of stem cells. What happens to their personality, to their memories? No one knows. This is not a philosophical oddity. It's actually very important in terms of what is going to happen during these therapies.
I want to talk about now - all of this plasticity is building up to the following: The idea that the body can change, the brain can change, the mind can change. To be appropriate, the different types of bodies. There's this incredible plasticity of self-repair, dealing with novelty, behavioral novelty but also anatomical novelty. These are the examples I've shown you. One thing that's definitely happening is that we are leaving the garden of Eden in an important sense through a science of chimeric bioengineering.
This is a famous painting of Adam naming the animals in the garden of Eden. At some point a couple centuries ago, what you could entertain was a world view in which there was a finite fixed number of animals. They were distinct and you can name them and that was it. This world view is crumbling because, as it turns out, we can readily - biology is very accessible to readily mixing contents of different creatures at different scales. You could makes DNA, mix cells, transplant organisms, mix organisms together and make hybrid storms.
People knew this already, they knew you could mix the genetic material of a donkey and a horse and get something different. We already had these kind of hybrids. Now with chimeric buy other engineering, the opportunity to make different novel creatures has arose in a different way. At any level of organization, at the level of molecules with Nano materials and smart materials, at the level of cells, individuals, colonies, we can replace any part of this with either organic, meaning evolved, or cybernetic or inorganic materials, we can replace any part of this. This is now a wide range of - in addition to the normal kinds of humans and other animals and so on that are out there, there are now humans with intelligent appliances.
These might be neurotransmitter pumps and other things. Humans with brain implants that control wheelchairs and other devices. Lots and lots of different types of creatures. Now I want to show you one example that we created. This is collaborative work led in my group Douglas Blackeston. My collaborator is Josh Barngard and his group.
We took skin cells, liberated them from a frog embryo. We asked the question, without any further manipulation, no genomic editing of any kind, freed from the constraints of the embryo, what would they build, if anything? What would they do? What would these cells do? If they a chance to division their multicellularity. They don't crawl off, don't make a flat mono layer.
They compact together and then they start this very interesting activity I'm going to show you. Once they create this kind of body, they become what we call xenobots. So xenobot is from two words, xenopus laevis and from the frog, and bio bot. The way it's swimming along, it has hairs, cilia on the outer service. Normally the cilia are there to redistribute mucous over the skin of the frog or the tadpole.
You can see they go in circles, they can go straight back and forth. There's one patrolling. Here is tracking data on a number of them. These are interacting with each other. This one is going on a circle, this one in a long trip around the petri dish, and this one is doing nothing.
One of the ways I sometimes do this talk is I show a lot of o videos what they're doing and I don't tell people what this is and I ask them to guess. You would never guess this is a bunch of frog skin that has decided to form a new kind of proto creature. Here is what they can do in a maze.
This is a water-filled maze. There's no flow. They're swimming along. It takes the corner without having to bump into the wall. Over here, internal dynamics kicks anywhere it decides to turn around and go back where it came from.
We'll run it again. It goes forward, takes the corner, doesn't bump into anything and then turns around and comes back. They have spontaneous behaviors. Amazingly they can regenerate back to their original shape. Here is a xenobot that was basically cut in half like this, and that hinge generates massive amounts of force through 180 degrees to pull the whole thing back into a more proper xenobot shape.
Not only that, if we look at calcium signaling which is basically a readout for electrical activity, and here are two setting next to each other, we can see they're extremely active. This is basically what a lot of brains - a zebra fish brain would look like. It's brain-like activity.
Maybe they're in communication. We don't know that yet. This is just skin. The incredible plasticity of the cells liberated from their normal environment to give rise to a proto animal that can reuse its various genome-specified hardware features to - in novel ways, in new circumstances as a new creature.
Why do we want to make these? First of all, there are applications for useful synthetic living machines. If we can learn to control them, maybe they are scraping the plaque off your arteries or fixing up arthritic knee joints or finding tumor cells in your gut or bacteria. So lots of synthetic living machines. More importantly, I think, the system provides a kind of sand box for cracking the morphogenetic code, what is the software, what are the algorithms that allow these cells to join together in a new way, in a novel environment.
So really to understand why is their shape the way it is, their behavior the way it is? That gets us closer to the idea of an anatomical compiler. We now understand how cellular collectives make decisions. We can impact those decisions and grow a new heart or new leg. The bigger philosophical question is this, where do their goals come from? In other words, they have primitive goals in the anatomical sense, meaning they make a specific functional structure. They have certain behavioral repertories. For millions of years, the ancestors were selected for the ability to do this or that.
This is just skin, selected to sit quietly on the outside of a tadpole and keep the pathogens out. If you sequence the genome, all you ever see is frog. You don't see anything other than completely standard - where do the goals of this novel collective intelligence - again it's the collective intelligence because it's the product of lots of cells working together to do this, is critical. Modern science and engineering is making tons of new collective intelligences. These are called kilo bots and they can be programmed.
They'll do things - you can't predict them without knowing the rules of each individual bot. This is a kind of existential question for us going forward in society, you make internet of things and all these swarm robotics, what is the swarm going to try to do? What are the goals? Where do goals come from? It's a major unanswered question. The ability to make these kind of things and observe the novel behaviors and structures that appear, are an important way for us to close this knowledge gap. Just to finish up here with a couple of points, everything about biots, as I pointed out is swappable. Here is a system where the normal structure of brain, body, environment has been altered. Here is a brain.
This happens to be a dish with some rat neurons in it. But the signals of these neurons are processed by computer and fed into a flight simulator game. These cells are rewarded or punished based on how well the flight does. If the plane crashes, they get electric shock and so on. Sure enough, the neuro cells learn to fly the plane. This is now their body, this is what they're able to do.
That plasticity, the ability to hybridize basically with virtual bodies, virtual environments is really important for the following reason. Here is an example. Here is a gentleman that has no biological arms, but he now has robotic arms that very much like the neurons you just saw flying the plane, now his neurons are able to drive these mechanical arms. So the plasticity of these - these are called Hybrots, a set of neurons used to drive a neural body. Their body is a vehicle. This is telling us something very important.
This forms a continuum. Here you have a human with a little bit of electronics in. Let's say 95% human, 5% electronics. You would say that's still a human being. Here you have the opposite situation. You have a robot with a few human cells cultured on top.
You would say that's 95% robotics and 5% human cells. That's basically a machine. The amazing thing is we can fill in any point along this continuum. If you wanted something that was 50/50 or 70/30, it can be filled in by a possible human being. This idea it's a human, a robot, a machine, this binary sense that we can make a sharp category simply doesn't exist anymore. When one considers the true diversity of possible agents, you can imagine all the synthetics and the cyborgs and all these different things, it completely erodes these binary kinds of categories that we used to have.
In the olden day, you could knock on something and hear a metallic sound, you'd say that's a machine that came out of a factory, will perform in a predictably boring way. Whereas, if you do that and it's soft and squishy, you'd say that's an evolved creature and I have to take care of it. That distinction is being eroded.
What that means is, in the future, when we look at creatures in our environment and consider what moral and ethical responsibility we have for them, we will not be able to make those decisions based on what they look like and where they came from. And this I think resonates with a lot of other progressive ideas where you simply can't decide what you owe something based on its origin story, whether it got here by evolution or design or some mixture of the two, and what it looks like. We will have to be very, very cognizant of what kinds of minds these types of organisms have. Darwin at one point looked out into a river bank and he said - he was impressed by these endless forms most beautiful.
I want to point out that's basically - everything he saw in the living world, everything we see in the bio sphere is a tiny corner of this astronomical option space of creatures that could exist and are going to exist in our world. We're going to be living side by side with all these different combinations of biological structures, designed machines and software. All of these can be recombined readily.
Our strategy now is that we cannot fall back on this kind of simple criterion that people use in bioethics, in terms of - well, how much like a human brain is it? That's how we decide what moral protection something has. Is it an invertebrate, a worm, a dog, what do we owe it? These categories are being blown up by the realization that biology is endlessly plastic and recombinable, and we need to develop a much deeper understanding of what it means to be a cognitive being, a machine. We are all machines in a particular sense, and so on, and what we owe the various different types of creatures.
I want to close by summarizing what I've said. There's new biology that allows us to make all kinds of unusual new creatures - it's not about the unknown creatures, it's forcing us to hold up a mirror to ourselves, confronting profound issues that have always been here, but obscured by the familiarity of standard animal and plant models. The fact that zebra fish always make zebra fish. These have ethical implications well beyond the simple questions of bio safety. The familiar categories upon which ethics theories are built are basically crumbling. The distinction between robot, machine, organism, they don't exist the way they used to.
They need to be replaced. The young people in the audience will be living amongst a huge diversity of creatures with novel bodies and novel kinds of minds. Their origin and physical appearance will not be reliable guides to the moral responsibilities that we owe them. We need to develop new frameworks that simply don't exist yet. Everything of interest to ethical considerations including memory, cognition, consciousness and so on, are on a continuum. We must take evolution seriously, and whatever capacity a human might have that deserves certain kinds of moral considerations and responsibilities, you can trace that very smoothly all the way down to bacteria and beyond in increments that do not permit any kind of a sharp line with respect to almost anything interesting.
Evolution optimizes quantity, not quality. I would argue we have an imperative to improve the status quo. I think surely we can do better than random search. More details here if anybody is interested in fuller treatments of all of these questions. Here are some papers you might be interested in.
I want to thank the people who did a lot of this work. These are the post docs and students in the lab. I want to thank my collaborators, people who fund the research we do, of course, all the animals and model systems we work with. I thank you very much for listening. I'll take questions.
>> JONATHAN BEEVER: Thank you, Dr. Levin. That was great. We call all give him a virtual round of applause in that awkward Zoomy kind of way. We've got a lot of questions. I notice, Dr. Levin, the questions are in two forms.
I want to start with what I'll call the technical and research ethics questions and expand out into the bigger-picture questions. Folks can use the Q&A top adjust their questions if I'm misrepresenting anybody. I'll start with a couple of these technical questions. The first question was how did your team determine what amount of signaling was needed to regenerate a frog's limb? >> MICHAEL LEVIN: Basically what we've been doing for years is trying to understand what kinds of signals drive the creation of body structures in the first place.
We observe embryonic development, all the signaling that goes on so we can track the electrical conversations, the chemical signals, we build computer models to see how do the conversations scale up to a larger type of goal. No individual cell knows what a leg or finger is, but the collective does. So what we found were some triggers, bio electric signals that basically triggered to the cells build the particular structure here. We can make eyes, limbs, hearts and a couple of other things.
There's many things we don't know how to induce yet. That's where it comes from, from a study of embryonic development. >> JONATHAN BEEVER: That's helpful. Relatedly, another question about whether the frog's skin cells that create xenobots continue to lead to the same organisms, or is there a substantial differentiation every time you run the process? The question is do you end up with differently structured xenobots, each iteration? >> MICHAEL LEVIN: Right.
It's interesting. It's much like with any other aspect of biology, the large scale is very consistent. So most xenobots look basically exactly the same at the large scale. But they all have interesting individuality. There's motion types, a huge range of different motion types, a huge range of different responses to different behaviors.
So lots of individuality. That body type is a particular thing these cells like to make and it's quite consistent. >> JONATHAN BEEVER: Okay.
This next question comes from Professor of Chemistry Eloy Hernandez at UCF who asks how can you explain the apparent contradiction between evolution and the self-corrective process that you showed in the Picasso's tadpole example? >> MICHAEL LEVIN: I didn't hear the second part of the question. >> JONATHAN BEEVER: Self-corrective process that you showed in Picasso's tadpole example. >> MICHAEL LEVIN: The idea is this, you can think about the process of evolution. The most mysterious thing about it, certainly as I heard about it as a kid, it seemed wildly implausible.
We all know if you're building something or writing computer code, random changes are much likely to make things worse, not better. The question is how in a reasonable time frame do you get all these remarkable capacities that we see in the biological world. What I think is absolutely required for that to actually work is an important degree of what we call multiscale competency, meaning that if the eye is off of where it needs to be, it is able, within limits, to sort of on its own, without any genetic instructions to the contrary, it's able to find where it needs to go. That's part of what makes evolution work. It means, if you had animals who are hard coded - and there are a few that - any time you had a mutation that would change something, everything else would fall to pieces.
If you had a mutation that moves the eye a little bit but maybe does something beneficial somewhere else, that animal would be dead. The eyes wouldn't work. The mouth wouldn't be in the right place, the animal would be dead and evolution couldn't reap the positive benefits of that mutation anywhere else. The fact that these individual tissues and organs are self-correcting in anatomical space and physiological space, transcriptional space, they have a lot of capacity to make up for differences in their environment.
Some amazing examples I could give you where that means it acts as a buffer, means that evolution can explore all kinds of potential changes because things will try to get their job done even when their environment is different. If the eye ends up on the tail, no big deal. The optic nerve still knows how to find the closest large neuro fiber and synapse onto it.
If information comes from the spinal cord, I can still use that. It doesn't have to come through the optic nerve. That plasticity I think is a crucial part of evolvability.
It's what makes the process so effective and rapid. >> JONATHAN BEEVER: Relatedly, John Weishampel, professor of biology here asks the question - he says fascinating research on understanding fundamental biology which may have myriad after applications, any thoughts on the mass extinction of species that evolved over the millennia versus the creation of novel organisms? >> MICHAEL LEVIN: That's a great question. >> JONATHAN BEEVER: Now we're going bigger question.
>> MICHAEL LEVIN: I don't have too much to say about extinction, but I will say one of the biggest issues with open questions of evolution is the so-called arrival of the fittest. I think we understand selection fairly well, and you can sort of see how the environment and competition would winnow out things that are suboptimal. But where do body plans come from in the first place is a really interesting question.
One of the things that I think the work, for example, in the xenobots is telling us is that there are aspects of this that we simply do not understand. Unlike every other creature on earth that I'm aware of, these things don't have a lengthy evolutionary backstory of selection towards their shape. This thing basically put itself together in 48 hours in front of our eyes under completely novel circumstances that in this lineage did not exist.
So the ability of - this novelty, the ability of, if something happens to the genome and now the mechanism is telling the system where the signaling molecule for where the eyes are supposed to go is in the wrong place or whatever, all of these are viable organisms. That's the amazing part about this. The frogs with the eyes on their tails, six-legged frogs, two-headed worms, all of these are completely viable. They may not be ecologically competitive with others, but that's okay. They're viable creatures. There's this incredible variety of variability of living things.
That's important for the appearance of new species. In fact, something I didn't have time to mention, but those two-headed worms are, in fact, a kind of new species. If you cut them into pieces again, they continue to make two-headed animals.
Even though their genomes are completely normal, no changes to the genetics, the question of how many heads you have is actually stored in an electrical circuit that we've modified to make two-headed animals, that circuit has its own memory. They'll continue to make two-headed worms in perpetuity as far as we can tell. Is that the creation of a new species? I don't know. Much like everything else I've shown you, it kind of erodes these classical definitions. It's a permanent new line of animal. >> JONATHAN BEEVER: My brain is thinking about connections to the extinction debate and whether bringing back an extinct animal is the same thing as creating a new species.
Super interesting. We've got maybe two more questions about the big picture biology things. Then I'll ask you a couple of questions big picture about ethics specifically. This comes from Mason Cash, a professor of philosophy who says you said there are no bright lines between different levels of cognitive sophistication.
But many argue there are really bright lines especially in cognition, responsibility, self-awareness caused by environmental factors like language. A human child raised in a social world without human language, raised by wolves, lacking a history of development in that cognitively enhanced environment ends up with a dramatically different cognitive capacity. So development of cognitive capacity relies on strong environmental factors like language that scaffold cognitive development. Languages are ratcheted to help lock in learning and changes. So the question is are you understanding the ways biological and evolutionary sculpted development has been fine-tuned by developing and/or evolving in a context that includes the requisite environmental scaffolding? Would it be possible to replicate that kind of development? >> MICHAEL LEVIN: It's a good point. I don't disagree with anything you said in the sense that, What I certainly don't mean is that there aren't great transitions.
There are qualitative leaps in cognitive capacity. I showed you one possible ladder. I totally agree that language is such and so on. What I mean by no bright lines is that people often, when they're not careful, they often talk about these things as if they were binary categories.
The reason why that doesn't work is because what that asks you - certainly creatures with symbolic language are different than those without. But what you're not going to find is a single point in evolution where, here were some parents that did not have feature X and then, bang, their offspring - here is the change and their offspring have full-blown language. We don't find that.
Even though certain cognitive capacities might have evolved rapidly, rapidly is only rapid on the geological sometime scale. If you look at the biology, many, many generations it takes. There's nowhere where you can put a binary - that creature has it - doesn't have it and then the offspring do. So that's all that I'm trying to put an emphasis on, that all of these things are - the biology is changing quite slowly, and the result is often - again, it only looks rapid on a geological time scale actually. All of the changes are quite piecemeal.
That's why I think we have to be really careful, because whatever kinds of - let's say responsibilities or, in fact, privileges that we give to a particular kind of creature, we have to ask what do we think about very, very similar creatures which we can make or which already exist that are all in the space of that adjacent possible. We don't - and that's really the thing, is that all of these things are not crisp categories. They're all a cloud of other creatures that are similar to them in some ways, different in others, whether they've been modified or evolutionarily different. >> JONATHAN BEEVER: It seems like the sorts of questions and responses you're working with in terms of artificial organisms or artificial life mirror the same sort of work that's going on across lots of other species.
There's all kinds of interesting work on cephalopods and intelligence and cognition. Same with like tree communication through chemical responses to predation, and all sorts of things. Sort of seeing cognition, or whatever you'd like to call it, across this gray scale or model. It seems like you're coming to those same questions through the work on xenobots that other folks are coming to working on existing organisms. >> MICHAEL LEVIN: I think that's true. So evolution is a slower version of what's being done in synthetic bioengineering.
Evolution has been telling us forever that these creatures are part of a slowly changing continuous lineage, right? We are just expanding that into - much more rapidly expanding it into multiple directions across the space of all possible creatures. >> JONATHAN BEEVER: Okay. The last of the biology questions comes from Professor Will Crampton in biology who asks, Physicist Richard Feynman once said we can't make what we don't understand.
One things humans have never managed to do is make something living out of something dead. Your wonderful artificial xenobots are made from the cells a frog while Craig Venter's synthetic mycoplasma laboratorium was made by injecting a man-made genome into a bacterium stripped of its own genome. In your opinion is it conceivable that humans at least in our lifetimes could create a truly synthetic self-replicating organism from scratch, one without parts made from other living organisms? >> MICHAEL LEVIN: Yeah. Well, the simple answer I would say is yes.
I think that's absolutely possible. The deeper issue here is that, much like with many other terms that have been eroded by modern progress, I'm not exactly sure what living means anymore. I'm not alone.
There are lots of people that work in exobiology and try to understand say, okay, if we were to examine other worlds, what is it that we're actually looking for to say, ah, that's life? There's lots of really interesting work on what's called meta materials and smart materials. That - given that all of life is also a continuum, the bottom end of that continuum from bacteria and sort of lower down - I think melds very, very continuously into material science. So figuring out where you'll put a line - I've heard this before and it drives me crazy, well, that's just physics. Guess what? Everything is physics. There are different kinds of physics exploited in different ways.
I think there's going to be a smooth gradation between things we always recognized as living and things that look like just physics. There are amazing examples of oil droplets that also solve mazes. I think those two fields are going to merge. >> JONATHAN BEEVER: We're short on time. I want to ask two last question to wrap up.
One is very big picture. This comes from Rabbi Sanford Olshansky who writes in the chat, you made no mention of religion from which most American western ethics is derived. What should be the role of religion in determining the ethics of the modifications of typical life forms and the creation of new artificial life forms? >> MICHAEL LEVIN: That's not a small question. Okay.
I'll take a stab at it. So this is my belief, that whatever the role of religion is going to be, it absolutely has to be tied closely to what we understand about the natural world. I think to the extent that religion is giving us useful frameworks for action, it has to be consistent with what we know. And that's fine. As far as I can see, there are many compatible ways to understand religious precepts in ways that are compatible with the science. But I think all of it has to work together towards a kind of life-positive outcome.
How ever we interpret religious thought, I think it has to be in a way that is going to be for the benefit of society, and that means, to my way of thinking, that means it has to be firmly grounded in the reality of what we know. So if there were religions that were to say - this is actually very interesting. I, in fact, have an appointment next week to talk to the rabbi at Tufts about similar issues. I've worked with some Buddhists and so on, about the status of these various creatures. Actually, if someone had a religion that said there are only X number of creatures and that's all that there are, that is clearly incorrect.
If we're going to have useful religious framework to guide our activity, it has to be brought up to date with what we now see as possible. We have to take these creatures seriously. We have to honor their ability to suffer and we have to modernize as much as necessary those ideas. >> JONATHAN BEEVER: Thank you.
I know that was a big question. I want to circle back around, I wish we had more time, to what I took to be your central thesis, which is we've actually got a moral duty to improve the world through this type of research. >> MICHAEL LEVIN: I think so. >> JONATHAN BEEVER: As opposed to a risk-based view of technological development and scientific progress. But my mind - when you stated your thesis, my mind immediately went to all the litany of ethical missteps that we've taken in scientific development, both on biology end, but also on information systems and destructive systems, and you can go on and on. I wonder if you could talk with us briefly again about that central thesis and how you see that - I don't know - risk-benefit analysis playing out.
>> MICHAEL LEVIN: I think it's for sure true you have to do the risk-benefit analysis as much as possible. It's hard because we often don't see all the implications of our work, for any specific case. I think it's essential to think about the possible misuses on all of this stuff. If we take the 30,000-foot view and ask broadly how often is it the case that we would have been better off without scientific inquiry? I suppose this could be debated. I'll give you my personal view. It seems to me completely convincing that overall science is a good thing.
Much like anything else, we screw it up from time to time and misuse it in various ways. But overall, I can't think of a pre-scientific time period - in fact any time period prior to now that I would like to live in versus now. I know other people are more competent with me will have statistics about what it was like and how it was to live in these other areas in terms of violence, and just the ability - to me it's the ability to be more than the biology.
I think this gets back to the previous question. Survival is fine, but the ability to actually live in the world where you have free time and the capacity - the freedom and the capacity to sort of explore and exploit your creative gifts so you're not spending all your day digging for roots. How many tens of thousands of years of that do we really need? I think science lifts us up out of mere biology to sort of more exalted ways of being, and I think that's essential. I think we are going to mess it up from time to time, and we have to do our best.
But I don't see any way - Especially with the medical suffering in particular. People often will say to me, oh, that's too much. I'll say, well, pick a point. Because at some point putting a splint on a broken bone was maybe considered too much, right? Is that too much? Blood transfusions? I suppose some people think that's gone too far.
Now we have brain surgery, we have all these things. I don't see any moral way to tell any of the people to call me to say my kid has this problem, I've got a spinal cord injury, I've lost a limb, I don't see any moral way to tell any of these people that we've decided enough is enough, we're going to stop here. I think that's a complete non-starter in an ethical sense. >> JONATHAN BEEVER: Thank you very much.
There's endless questions we can talk about from concerns about justice and access to the sorts of technologies we're talking about. I'm grateful, Dr. Levin for your time and to everybody for asking really excellent questions. Thank you very much.
I'm going to turn things over to Dr. Kuebler to wrap things up. Thank you again. >> MICHAEL LEVIN: Thank you very much.
>> STEPHEN KUEBLER: Thank you Dr. Levin for an intriguing and thought-provoking presentation, and thank you participants for your outstanding questions. A link to this presentation will be made available at the UCF Center for Ethics website so you can view the recording again and share it with others.
And there you can also find information on up
2021-11-06