On The Future of Body, Mind and Intelligence
What is an Anatomic Compiler? What is multiscale cognition? How about bioelectric fields and cognitive glue? A mind-bending interview with developmental and synthetic biologist Dr. Michael Levin.
Michael Levin is a developmental and synthetic neurobiologist at Tufts University and, according to many, on track for a Nobel Prize for his work unpacking new ways organisms develop and function, with deep implications for regenerative medicine, for treatment of cancer and other conditions, and for consciousness and intelligence, human and beyond. His laboratory work demonstrates paradigmatic ways bioelectricity can store information and shape development, free from genetic influences.
In work with flatworms, for example, Levin and his team have been able to switch between producing one-headed and two-headed worms based only on whole-organism electrical fields. The forms life currently takes are highly constrained in “morphospace”. There is much more we can do as we come to understand the inner workings of biology, and computational reality more generally.
Levin’s lab has shown that frog embryos, while still visibly a ball of cells, have “electric faces” that can be visualized with voltage-sensitive dyes. And the researchers have suggested that developmental processes, far from following an exclusively genetic plan, can be viewed as problem-solving. They can make a “Picasso frog” by moving cells from their positions in the electric face — and work out together how to get back into the right pattern.
In other experiments, they have shown that individual cells, which do not show such behavior in nature, can form aggregates in a dish that behave like a new organism, swimming around and collecting food. Importantly, the cells can also recruit to their group other cells that will then show similar behavior. Generally, he calls these “biobots”, and his lab has shown that when they have neural cells in them, neurobots exhibit more complex behaviors. Separated from the body and its bioelectric environment, the cells act as autonomous selves. They can rejoin regular tissue and become part of the organism again.
Such “anthrobots”, made from human cells, may be useful for speeding up wound healing. Their activity is also related to how cancer works — forgetting, if you will, that they are part of a collective. Levin’s team has been able to induce genetically damaged cancer cells back to normal using bioelectric-field manipulations still early in development.
I had a chance to interview Dr. Levin a few months ago in person1, and he was kind enough to take some more time to answer bigger-picture questions about recent work he is publishing2 (Machines All the Way Up and Cognition All the Way Down: Updating the machine metaphor in biology). Here is our interview.
GHB: Let’s start big-picture. How does our view of consciousness and intelligence, as currently held, limit us?
ML: Those are two separate (but obviously related) issues. With respect to intelligence (publicly observable competency to reach goals by different means, in some problem space), there are two fundamental issues that limit progress (in science, engineering, and ethics).
a) Many people think it’s binary. That’s wrong, it’s a continuum, and the right question isn’t “is it intelligent or just physics”, but “how much and what kind of intelligence does it have”. It goes all the way to the bottom, with even simple molecular networks having learning ability and other mind-relevant competencies.
b) Many people think you can just decide, using ancient philosophical categories, what kinds of things can be intelligent — very limited view focused on brains and medium-sized objects moving at medium speeds through three-dimensional space. It’s hard for people to recognize intelligence in unfamiliar guises operating in unconventional problem spaces that biology exploits (for example, the sensing-decision-making-action loops that cells make in physiological and anatomical state spaces).
This mind-blindness (due to binary thinking and brain chauvinism) prevents us from deploying powerful tools from behavioral and neuroscience (which isn’t about neurons at all but about scaling up cognitive capacities in multiscale systems) in novel contexts, which, we have already seen, pays off in really powerful ways in terms of novel research capabilities and discoveries.
With respect to consciousness it’s hard to say much right now, except that the science shows us that for the exact same reasons that we attribute consciousness to each other (the problem of other minds), you should be take very seriously the possibility of (non-verbal) consciousness in your body organs, for example. People need to understand that we do not have a mature theory of why our proteinaceous brains are associated with consciousness and certainly can’t rule it out in other media.
GHB: What is an “Anatomic Compiler”? How might the way we conceptualize regenerative medicine, and more broadly our notions of bodies and physical form, change?
ML: The anatomical compiler is my vision of the future of the field. Some day, you will be able to sit down in front of a computer, and draw the animal, plant, organ, biobot, etc. that you want — any living structure, no matter how conventional (naturally evolved) or weird (designed by engineers). The anatomical compiler will convert that anatomical description into stimuli that have to be provided to cells to get them to build exactly that. If we have a way of controlling what exactly cells build, then almost all medical problems — birth defects, traumatic injury, cancer, degenerative disease, aging — go away, solved permanently.
The key, though, is that the anatomical compiler is not some 3D printer by which we micromanage the structure of organs — it’s a communications device — a translator by which we convert our goals as bioengineers (or workers in regenerative medicine) into the goals of the living matter (an agential material), using bioelectrical, biochemical, and biomechanical interfaces to communicate and collaborate with the native problem-solving, memory, and decision-making capacities of the collective intelligence of cell groups.
GHB: What are “bioelectric fields” and “cognitive glue”, and how do they hold promise for various applications?
ML: Bioelectric patterns are memories of the morphogenetic intelligence of cells (much like the electrical patterns in our brains encode memories, goals, and preferences) of the conventional behavioral intelligence. These patterns control what cells build — they specify the future shapes of organs and tissues during embryogenesis, regeneration, and cancer suppression. These are the patterns that we try to re-write when specifying new goals to the cell groups, which is how we get them to build eyes, limbs, repair birth defects, normalize cancer cells, etc. Bioelectric signals are one of the kinds of cognitive glue — a mechanism that binds competent parts (like cells, which also have the ability to learn, and pursue their own local agendas) into wholes that have larger minds projected into new problem spaces.
Bioelectricity has long been known to be the cognitive glue in the brain — ability of your neurons to communicate in networks is what enables you to know things that your neurons don’t know and to have goals in spaces that your neuron cells don’t operate in (social space, linguistic space, financial space, etc. not to mention 3D space). The same is true for the rest of the body because the brain’s cognitive glue evolved from the much more ancient use of electrical networks to bind individual cells into a collective that knows what an embryo is supposed to look like (even though no individual cell does).
The emerging science of bioelectricity, and the cognitive approach to it (i.e., it’s not just physics, it’s also decision-making and memory processes) unlock transformative approaches to regenerative medicine. When taking seriously the problem-solving competencies of cells and tissues in anatomical space, we don’t need to micromanage cells or genes — just like we don’t need to worry about rearranging the synaptic proteins in a person when we talk to them so that they understand us, we can communicate with the cellular collective and ask it to build complex structures when we have no idea how to micromanage it from the molecular/genetic level. We have shown many examples of this — communicating complex, top-down goals (like “make an eye”) to cells with simple signals and the cell groups manage all the downstream molecular biology themselves.
GHB: What is “multiscale cognition” and what problems does this conceptualization address? (e.g. breaking down duality) What are some of the pragmatic outcomes from this approach?
ML: Multiscale cognition is the discovery that in living matter, not just the animal but the collective (think: ant colonies) and the parts (organs, tissues, cells, and molecular networks inside cells) all have their own agendas — goal-seeking competencies and problem-solving skills in diverse spaces (as well as memory, predictive capacity, etc.). The fact that we are made of agential matter (not passive matter, like metal or plastic, nor even active matter or merely computational matter, as exploited in engineering now), has massive implications for evolution: Evolution works very differently on a competent material than it does on passive matter.
The property of material to problem-solve — from the most basic molecular level on up — helps us understand many areas, present and future: the intelligence ratchet of evolution, or how intelligence grows as organisms become more complex; new options for biomedical development — rather than just rearranging DNA and other molecules, directly talking to your organs/cells to transmit complex goal states and get your organ and cellular buy-in for stable cures (not just symptom relief while you’re taking a drug); and bioengineering advances, to create novel synthetic useful machines.
Finally there are key ethical considerations, given the conceptualization we are discussing. Learning to communicate with unconventional minds is mission-critical given the forthcoming plethora of cyborgs, hybrots, and other enhanced and chimeric “humans” I believe we can expect to see in the very near future.
GHB: How do you think about consciousness, AI and self?
That is a very deep question on which I’ve written probably hundreds of thousands of words 2. To summarize briefly, Selves are multi-scale systems that continuously interpret their own memories to guide adaptive action toward remembered goals; they bend the option space of their parts to expand their cognitive light cone into larger goals projected into new spaces. They can be made of many substrates — and have many possible origin stories — the random meanderings of the evolutionary process across the space of proteins has no monopoly on making real Selves.
I do not believe we make conscious beings — either by having babies or making “AI” — what we make are physical interfaces or pointers into a Platonic space of forms (some of those are simple, low-agency things like the truths of mathematics, and some are complex high-agency patterns that we would recognize as behavioral propensities, a.k.a., kinds of minds).
Biological bodies enable the ingression of familiar evolved forms of structure and function, but our recent work on the creation of synthetic multicellular life forms (using normal cells with un-edited DNA) which have amazing new shapes and behaviors (which cannot be explained by a history of selection, since they never existed before!) shows that there are different patterns that can be embodied from the same genetic hardware.
And of course chimeras, cyborgs, and AI will show us kinds of minds that may never have been embodied in the universe before (certainly not on Earth, at least). All of their amazing behaviors are what we see from third-person perspective, observing them in the physical world, but consciousness is what we call the point of view from the Platonic space looking out — because we are the patterns and our consciousness is the inner perspective of being a pattern with a temporary embodiment into the physical world.
GHB: What might the world look like in 100 years?
ML: I don’t think we can say what the world will look like in 5 years, nevermind 100 years… But let me paint you a picture of what I want it to look like (if we survive and mature as a species, we may get there, and I don’t think it will take 100 years). I see a world in which freedom of embodiment means that you are not subject to the whims of genetics or random mutations of history but can be intentional about how you exist in the physical world; where you can focus on fulfilling your creative potential not hampered by diseases or limitations of body or mind.
You have not only limitless healthy lifespan but also the ability to change your embodiment in whatever way facilitates your dreams, because we have full control of biological growth and form, and thus of our bodies and cognitive hardware. The people of the future will look back on us the way we look back on our hominid ancestors, wondering how they could possibly have lived in a world where falling on a sharp stick can mean sepsis and death. I imagine them saying, “can you believe that back then, you had to spend your whole life in whatever body you were given at birth — susceptible to horrible diseases, IQ ceilings, and other limitations determined by stray cosmic rays hitting cells which cared nothing about values or the meaning of their lives? Incredible that they could live like that.”
But more than that, this is not just about physical or technological change — it is about spiritual advances as a mature species. The point is that by recognizing the ways in which minds embody in multiscale systems, we can recognize and have an ethical synthbiosis with all kinds of diverse intelligences that exist un-recognized in our midst, from ecosystems to patterns of information and energy within our bodies and our environment. We will be able to expand our cognitive light cone to deploy active compassion much wider than our current firmware (with its emphasis on genetic relations, in-group and out-group conflicts, etc.) allows.
Michael Levin is a developmental and synthetic biologist at Tufts University, where he holds the Vannevar Bush Distinguished Professorship and directs the Allen Discovery Center. He is renowned for pioneering research in bioelectricity, regeneration, and among other mind-bending real-world applications, the creation of Xenobots — living robots made from frog cells, and the capacity to switch on and off the number of heads a flatworm has.
References
- Engineering Body and Mind with Dr. Michael Levin, Doorknob Comments #82
- Machines All the Way Up and Cognition All the Way Down: Updating the machine metaphor in biology
- Please see the definitions in items #13 and #14 at https://drmichaellevin.org/resources/.
Originally posted on Psychology Today, ExperiMentations
