The field of Artificial Life aims to study life as it could be1,2. One definition of life is based on autopoiesis, or self-making3,4. Living organisms such as humans continuously break down and re-make proteins, subcellular organelles, cells and tissues in their bodies, unlike any machine. Cells and organisms do this self-making inside themselves. It is an emergent property that can not be predicted based only on the properties of the individual components. Moreover, all living systems reproduce or perpetuate themselves by growth in or on the body, followed by splitting, budding, or birth5.

In contrast, subcellular organelles replicate without internal growth. Instead, they assemble material in their external environment, (but still inside the cell) to make functional copies. Prior to 2021, such kinematic replication had never been observed at higher levels of biological organization, nor was it known whether multicellular systems were even capable of it. In 2021, scientists from the USA changed this. They published a report of assemblies of heart and skin cells that replicated kinematically. They were made from stem cells from the African clawed frog Xenopus laevis. So, they are known as xenobots. They moved and compressed dissociated cells in their environment to make functional copies of themselves. This behavior arose spontaneously over days rather than evolving over millennia. In this study, methods based on artificial intelligence designed assemblies that replicated and performed useful work as a side effect of replication. Still, the authors called these reconfigurable organisms and machines5. In contrast, the theory of autopoiesis considers living organisms not to be machines and would not consider xenobots to be a form of life. Still, xenobots have been described as being living robots and a form of life6. Others have described them as self-propelled, autonomous proto-organisms that have some of the properties of living organisms, but not all of the properties (such as autopoiesis)7. Xenobots may have great value. They may become able to clean polluted oceans by collecting microplastics. Also, they may be used to enter confined or dangerous areas to scavenge toxins or radioactive materials. Xenobots might also be designed with carefully shaped pouches to carry drugs to target organs in human bodies7. They might also be able to remove cancer stem cells before they develop into mature cancerous cells, mature cancerous cells, and tumors8. They are less than 1 mm long and made of 500-1000 living cells. They can propel themselves, join together to act collectively, and move small objects9. Using their own cellular energy, they can live up to 10 days.

Artificial Intelligence (AI) and a supercomputer were used to test thousands of random designs of simple living things that could perform selected tasks5,10. After selecting the most promising designs, the virtual models were replaced with frog skin or heart cells, which were assembled manually. The heart cells enabled the assemblies to contract and relax, giving the xenobots motion. The cells were harvested from embryos at the blastula stage, when they mostly retain the ability to grow into any tissue type of the body. Then, selected functions were engineered into the design by using cell types that had begun to differentiate into a target tissue. Epithelial and cardiac cells were cultured separately. Next, they were combined so that they could aggregate into a single mass. The specific arrangement of cells was decided by using an evolutionary algorithm in simulations that looked for the geometries that performed best at the desired task. For example, a two-legged xenobot with contractile cells on its lower half was able to move in a non-random direction over a surface. The designs were optimized by iteration, using the best performing structures as the input for further rounds of testing. The motile structures also had emergent behaviors. For example, they could temporarily attach to or orbit one another when they collided. Some of the designs included unexpected features that might have new uses. One xenobot evolved a hole during the design stage that reduced hydrodynamic drag, and might function as a cavity for storing and transporting objects10.

So, whether they are called organisms or not, xenobots offer many exciting possibilities for improving human health and helping to clean the environment. Still, biologists might begin asking if autopoiesis (self-making) has to occur within a system for it to be considered to be alive. Can autopoiesis occur outside the organism?


1 Aguilar, W. et al. The past, present, and future of artificial life. Frontiers in Robotics and AI, volume 1(8), 2014.
2 Gershenson, C. & Cejkova, J. Artificial life: sustainable self-replicating systems. arxiv:2105.13971, 2021.
3 Capra, F. and Luisi, P.L. The Systems View of Life, Cambridge University Press, United Kingdom, 2014.
4 Luisi, P.L. The Santiago school. Autopoiesis and the biologics of life. Wall Street International, 29 Feb., 2016.
5 Kriegman, S. et al. Kinematic self-replication in reconfigurable organisms. Proceedings of the National Academy of Sciences, volume 118, article e2112672118, 2021.
6 Ramanujam, L. Xenobots: A remarkable combination of an Artificial Intelligence-based biological living robot. International Journal of Sociotechnology and Knowledge Development, Volume 14, 2022.
7 Abramson, C.I. & Levin, M. Behaviorist approaches to investigating memory and learning: A primer for synthetic biology and bioengineering. Communicative & Integrative Biology, Volume 14, p. 230-247, 2021.
8 Ghahfarokhi, M.H. Stem cell-based intervention of neurological cancer for cognitive restoration. Neurosphere Student Journal, Vol. 2, p. 11-19, 2022.
9 Brown, J. Team builds first living robots—that can reproduce. University of Vermont Communications, 29 Nov., 2021.
10 Ball, P. Living robots. Nature Materials, volume 19, March, 2020.