How We Accidentally Started Making Infinite Robots

Xenobots as they are called are synthetic life forms or living robots and consist of two things where skin cells and heart cells from the frog Xenopus laevis after which they are named where they are the first man-made organism and an entirely new type of living system and one day scientists hope that they will help us monitor radioactivity combat pollution or even cure disease demonstrating how synthetic biology creates novel biological systems by recombining cellular components in computer-designed configurations for biomedical and environmental applications.

Sometimes called soft robotics the field of biorobotics is a relatively new one where it’s a discipline that integrates biology with mechanical design where instead of using steel nuts and bolts the key raw material for biorobotics is living tissue where biohybrid robots are generally composed of muscle tissue either cardiac or skeletal and an artificial scaffold demonstrating how bioengineering creates hybrid mechanical-biological systems by replacing traditional materials with contractile biological actuators.

The benefit of using biology in robotics is substantial where biological components have unique characteristics that artificial materials can’t exactly replicate where their greater flexibility enables them to move more like living organisms than machines where they can also quickly and naturally respond to external stimuli where some could even heal from damage or injuries where artificial structures are prone to entropy the gradual decline into disorder but living organisms are in a constant battle against entropy where they preserve their internal order by taking energy from their surroundings demonstrating how biological materials surpass artificial counterparts through flexibility stimulus response self-repair and thermodynamic stability maintenance.

Scientists created a light controlled stingray where it’s a stingray shape made out of an artificial backbone that is layered with rat heart muscles where instead of being directly activated by electricity these muscle cells have been programmed to be activated by light where this way you can drive the robot around by shining lights on either the right or left hand side but for all these machines to work outside input is needed for them to know how to behave where electricity or light supplied by a human demonstrating how early biohybrid robots required external optogenetic control for locomotion unlike autonomous xenobots.

After a frog egg is fertilized it forms into a ball of stem cells where this clump of cells looks completely uniform but in reality all of these cells have a predetermined purpose a fate of what they’re meant to become where the cells on top generally become epidermis or part of the central nervous system the cells in the middle form the muscles and the cells on the bottom form the endoderm which leads to the development of organs where these different sections can be disassembled into their different components and then reassembled in a new arrangement where this disassembly and reconstruction is done entirely by hand using forceps where the desired sections of the embryo are removed then basically mushed back together where they are then bathed in a media that causes them to stick together and they re-adhere into a sphere then the spheres can be sculpted into new shapes using a cauterizer demonstrating how manual micromanipulation exploits embryonic cell fate specification to construct novel biological configurations.

A program called voxcad creates a virtual environment complete with real-life simulations of physics like gravity friction liquid physics and surface tension and in this environment are small cubes called voxels that represent the cells of a digital xenobot where different cubes represent different real-life cell identities where in the beginning the researchers started with just two cell types passive like skin cells and contractile like cardiac cells demonstrating how computational morphogenesis employs physics-based voxel modeling to simulate multicellular configurations before physical construction.

The different cell cubes are then combined in a random way by the AI and then placed into an evolutionary algorithm which can evolve the digital xenobot over time creating iteration after iteration until it displays the desired behavior where at first the algorithm makes completely random designs where these digital bots don’t exhibit any interesting behaviors beyond just sort of wiggling a bit but after many thousands of generations eventually the algorithm will produce a digital xenobot with the desired behavior moving forward demonstrating how genetic algorithms optimize morphological configurations through iterative fitness-based selection across thousands of computational generations.

The final design of the first digital xenobot was a little blob with leg-like appendages that it could use to scuttle in a walking-like motion where the scientists then make the xenobot in the lab with the same configuration sticking the correct cells together and sculpting it into the right shape and sure enough when placed in their aquatic petri dish the bots started walking and when placed in an orientation facing to the right they generally always walked to the right where as hoped the AI quite accurately predicted the way that the bots would move demonstrating how evolutionary computation produces functional morphologies that successfully transfer from simulation to physical reality with preserved locomotion behaviors.

The pink lines show the movement that the AI predicted and the blue lines show the actual movement of the xenobots where as hoped the AI quite accurately predicted the way that the bots would move where this predictable movement demonstrates that computer simulation successfully models real world biological behavior and that digital designs transfer to physical reality demonstrating how physics-based voxel simulations achieve sufficient fidelity to predict multicellular locomotion trajectories with minimal reality gap.

Scientists put the xenobots to the test to see if they could move through varied environments in mazes ranging from wide open fields to tubes as small as half a nanometer in diameter smaller than the diameter of a human capillary where in all cases the xenobots followed the path provided by the environment where traditional robots would have a lot of trouble navigating something this small but the xenobots could do so with ease demonstrating how biological flexibility enables passive microscale navigation through confined geometries impossible for rigid traditional robots.

The scientists also noticed that if they injured the xenobots after about 10 minutes the wounds would close and the bots would heal up where researchers are still working to understand the intricacies of this but it appears that self-healing is a built-in feature of these biobots where contraction at the wound site can help close the injuries demonstrating how biological damage response mechanisms enable autonomous tissue repair without external intervention through actomyosin-mediated wound closure.

Without being programmed and without any sense or communication organs the xenobots spontaneously started to work together collecting and organizing piles of debris where how is such cooperation possible by mindless little bots where a hint comes from a study done in 1999 where sensorless robots each with a little scoop simply drive around and by chance eventually start pushing discs where the robot can push one or maybe two but if it comes across a pile the robot isn’t strong enough to keep pushing so it’s forced to let go inadvertently adding its disc to the pile where over time piles spontaneously emerge where this is precisely what’s happening with the xenobots where they simply push particles by chance until they cannot which spontaneously leads to organized piles where even though it’s a mindless and uncoordinated act it could be a powerful tool demonstrating how emergent aggregation behavior arises from simple stigmergic interactions without centralized control or sensory feedback.

Scientists think this gathering behavior could allow such bots to collect microplastics from ocean water or to clear plaques from arteries in the human body and along with clearing pollution one day they could perhaps be programmed to selectively pick up and move specific cell types that we want to use in regenerative medicine aggregating the materials for a body to regrow damaged tissues demonstrating how emergent aggregation behaviors enable diverse biomedical and environmental applications from ocean remediation to cardiovascular therapy to tissue engineering.

First the scientists had the AI redesign the xenobots to optimize for collecting particles where the original sphere shape as you might guess isn’t so great for the collection task where instead the computer suggested a C-shape similar to Pac-Man which is highly efficient at collecting loose particles where then the researchers added a compelling material to the xenobot’s environment frog stem cells the raw material of a xenobot and as the scientists hoped the xenobots dutifully swept them up into small piles demonstrating how evolutionary algorithms optimize functional morphology to maximize particle aggregation efficiency through geometric redesign.

Then something remarkable happened where the piles of cells turned into new xenobots where there are many types of reproduction in the world from sexual to asexual from splitting to budding to birth but all come from the parent organism one way or another but what the xenobots did has never before been observed in living organisms where it’s called kinematic self-replication and has only ever been observed in certain molecules where it’s the act of reproducing by moving and compressing dissociated parts in the environment and although it’s never been observed in cellular life forms until now it’s possible it even played a role in the origin of multicellular life on Earth and this ability in robotics could lead to exponential utility over time demonstrating how mechanical aggregation of dissociated cellular components creates unprecedented biological reproduction mechanism potentially recapitulating early multicellular evolution.

Currently individual xenobots can live for 10 days in an aqueous environment without needing any outside food source where this is a good start but limits what a xenobot could do in the real world but if xenobot raw material could be continually added to the environment it could mean limitless generations of xenobots which could be a real force for good demonstrating how metabolic reserves from embryonic lipid yolk sustain autonomous function for limited timeframes but kinematic self-replication enables indefinite operational deployment through generational succession.

Researchers are now working to create xenobots that have a working memory a read write ability to record one bit of information using a fluorescent marker that indicates if they’ve experienced something in their environment where this type of memory could help us detect the presence of things like radioactive contamination chemical pollutants drugs or certain diseases demonstrating how biological computation through fluorescent protein expression enables xenobots to function as environmental biosensors recording exposure history for diagnostic and monitoring applications.

By understanding xenobots we start to understand the plasticity of living cells where freed of their evolutionary fate cells could be capable of astounding things where they can walk they can swim they can collect debris and reproduce where the question is what else might they be able to do demonstrating how cellular potential extends far beyond genomic programming when cells are liberated from developmental constraints and how novel morphological configurations reveal hidden functional capabilities.

As for whether or not these bots are alive is still a matter of great debate where they are made out of living cells and can reproduce but are still very much machines where our minds like to try to categorize but in reality we might need to accept that xenobots occupy a space in between both living and machine where for now xenobots are simple and lack nervous systems but if future biobots are capable of emotions feelings or pain this quandary may need more serious consideration demonstrating how synthetic organisms challenge binary life-machine classifications and raise bioethical questions about consciousness suffering and moral status of engineered biological systems.

Harnessing the power of biology will be the driving force behind the next century of technological developments whether it’s understanding the raw ability of cells free of their evolutionary fate or by analyzing the products of 3.8 billion years of evolution where the natural world has the answers we need in our quest to beat cancer find solutions to antibiotic resistance respond to natural disasters or traverse the toughest climates imaginable where our survival in the future relies on tapping in to nature’s genius demonstrating how biomimetic engineering and synthetic biology harness evolutionary optimization principles to solve critical challenges in medicine environmental resilience and exploration.