Researchers have produced revolutionary bio-robots that are capable of performing many tasks … and even replicating themselves.
After years of speculation on the concrete feasibility of such a project, scientists have finally managed to endow robots with an absolutely remarkable capacity: a form of reproduction!
This work was carried out by researchers from the universities of Vermont, Tufts, and especially from the institute Wyss, a Harvard laboratory specializing in engineering inspired by life. The latter are regularly illustrated with works that are often very original and quite impressive, both technically and in terms of their implications. And these “xenobots” are no exception.
These machines are remarkable right down to their origins, since they were developed on the basis of blueprints generated by artificial intelligence. Instead of being made of silicon, these little robots are made of stem cells collected from embryos of frogs. These cells are then conscientiously arranged in accordance with one of the billion models generated by AI.
This is important because these different stem cells will then differentiate into several cell types with very distinct roles. In this experiment, there are two particularly important cell types. On the one hand, we have epithelial cells that serve as a rigid support. On the other hand, heart cells act as a propellant by contracting.
Since the exact shape of these small robots is predetermined during a simulation, researchers can therefore produce very many varieties, suitable for various functions. At present, researchers have already produced xenobots that “swim, walk, push gravel, carry loads”… They can also accomplish more complexes and coordinates. For example, they can gather in swarms to “sweep” a surface by gathering all the waste in one place. Better: they are even capable of self-regenerate and record “memories”.
Last March, their designers passed a very important milestone by asking themselves a strange question; “What would happen if cells were “freed” from all their structural constraints?“Or in other words, to use the phrase of biologist Michael Levin:”what would these cells build if given the opportunity to re-invent multicellularity ?”
They therefore tried different approaches to let these cells assemble spontaneously. The result was a second generation of xenobots able to train on their own, without human assistance or guidance. Impressive progress, but still a long way from the feat of the third generation; these are indeed capable of reproducing!
The researchers had already observed that these organisms were able to replicate. But so far, all the baby xenobots produced this way have died fairly quickly. They therefore brought out their artificial intelligence to test a few billion additional shapes. But this time, they had a very clear idea of what they wanted to achieve: a xenobot capable of producing a stable offspring.
In the end, their system came up with an absolutely unbelievably behaving assembly. With his Pac-Man airs, this xenobot roams his Petri dish in search of isolated cells. Once he has collected enough, he will compact in a single structure which will then give a new xenobot, and so on!
A new functional conception of living things
It is an observation as exceptional as it is profound in its implications. Because in the end, this object is not at all a kind of chimera; the set contains exclusively the natural genome of the frog, but freed from structural constraints that would otherwise have forced these cells into a tadpole.
“It’s fundamentally different from how frogs usually do”, Breathes Sam Kriegman, lead author of the study. “No known animal or plant replicates this way”. This fundamental difference in operation has left the researchers. “We discovered this completely unknown space in living systems”Says biologist Josh Bongard. “We have Xenoboyts swimming, walking, and now spawning. The question now is what else to discover in this space.”
And from there, the theoretical possibilities are limitless. Researchers now hope to be able to improve their AI. the goal would be to be able very quickly produce a xenobot capable of performing a specific biological function, like “deploy these living machines to extract microplastics from a stream, or produce new drugs”, Explains Bongard.
And the potential implications could go even further. “If we can get them to do precisely what they want, at the end of the day, that is the ultimate solution in regenerative medicine. It would be the solution to serious trauma, birth defects, cancer, or even aging.”Enthuses Levin. “All of these problems exist because we do not know how to predict and control which groups of cells will form. Xenobots are a new platform to learn more about this.”
The text of the study is available here.