by Simon Coghlan, Ph.D. and Kobi Leins, Ph.D.
They’re neither a traditional robot nor a known species of animal. [They’re] a new class of artefact: a living, “programmable organism”.
These “reconfigurable organisms” have already provoked philosophical and ethical questions.
Xenobots are under 1mm in size and composed of 500-1000 cells from frog (Xenopus laevis) embryos. After culturing extracted embryonic stem cells, micro-surgery tools help “glue” the naturally sticky cells together in a range of simple organismic configurations, e.g. with molar tooth-shaped or kidney-shaped appearances. By then, the stem cells have become skin cells, which provide scaffolding, and heart cells, which spontaneously contract, giving the xenobots locomotion. Their design determines their autonomous functional capacities, which include being able, in aqua, to move linearly or circularly, swarm, “explore”, push small objects, and hold objects in a “pouch”.
Prior to biological assembly, in silica design models are generated by a supercomputer with an artificial intelligence (AI) algorithm. Primed with relevant biophysical principles, the “evolutionary algorithm” predicts which of innumerable possible cellular configurations will produce desired functions. In vivo testing results can then be looped back into the algorithm to generate new bespoke design models.
The biological and computational scientists who created xenobots acknowledge both possible benefits and potential ethical issues. To better understand these ethical issues, we should be aware of technological context and ontological considerations. Take technological context first. In synthetic biology (SynBio), scientists have engineered functional E. Coli bacteria with entirely human-made genomes that have much larger strands of DNA than in earlier efforts. Genomes in microbes can be chemically engineered to produce drugs, biofuels, and now nutritious proteins, raising the promise of replacing environmentally damaging agriculture.
In robotics, recent cyborg animals merge inorganic and living materials – such as the swimming robotic stingray made of rubber, gold, and rat cardiac cells, or the dragonfly with a backpack containing optrodes and circuitry that could turn it into a controllable cybernetic drone. In digital health, AI can, by crunching mountains of information, design pharmaceuticals far more rapidly than traditional methods. In nanotechnology, Trojan-horse carbon nanotubes have recently been designed to enter the bloodstream and attack atherosclerosis.
These fields and their strange discoveries may bring deeper insight into the mechanisms of life. Similarly, xenobots may facilitate biological understanding, perhaps by illuminating the emergent properties of living cells in their co-interactions. A distinctive feature of these fields is the increasing role not just of computing, but of rapidly advancing artificial intelligence. Though of mixed parentage, then, xenobots are partly the offspring of AI.
The above fields may also progress human health and wellbeing. Similarly, xenobots – or, if non-Xenopus cells are used, “biobots” – may be employed, for example, to enter living bodies to deliver drugs, detect and destroy cancer, and, like the nanoparticles mentioned above, remove atherosclerosis. When made from the patient’s own stem cells, biobots might be immunologically safer than other interventions. The xenobot scientists have also suggested that they hold potential for removing microplastics from the ocean and toxins from dangerous environments.
Now consider ontology. Being composed entirely of cells, xenobots are, in fact, not robots or machines in any traditional sense. So, are xenobots “organisms”? This looks puzzling. The glued-together cells survive for a week or so on embryonic energy reserves, although they live longer in nutrient-rich liquid. Xenobots function (e.g. swim) as an autonomous unit or assemblage. Their cells self-repair and reattach when cut.
However, xenobots cannot reproduce. More importantly for the ontological question, xenobots apparently lack the corporeally integrated and more extensive autopoietic (self-maintaining) functions characteristic of organisms. Arguably, then, they are “life-forms” but not organisms. Assuming major advances in AI and biology, though, that might change. Suppose, for instance, that future biobots were to become more biologically integrated, complex, and even capable of reproduction.
Although such possibilities remain speculative, imagining and pre-empting possible future technological developments is ethically important. Accordingly, we shall raise several ethical issues. First, future biobots may pose risks. Potential risks include dual usage possibilities, such as adaptation for use in warfare and assassinations, via the carriage and delivery of harmful agents. Historically, scientific advances have led to the development and use of chemical and biological agents in contravention of international prohibited weapons conventions.
Technology, including SynBio 3D printing, could conceivably enable “garage” biobot production. Accidental or deliberate release of more extensively autopoietic, long-lived, and/or reproductively equipped biobots might be environmentally damaging. Biobots could, for example, interfere with the life cycles of natural organisms, both in the sea, and, should they be engineered from cells from terrestrial organisms, on land.
Second, biobots with nervous systems might at a certain point become sentient. Sentience is often regarded as a characteristic that confers upon organisms some intrinsic moral status. In 2019, commentators raised ethical questions about the possibility of consciousness in a “revived” disembodied pig brain. The hypothesis that insects may have subjective experience is gaining traction. Although claims such as these are contestable, the possibility of sentience arising in relatively simple living things cannot just be ignored. Furthermore, some people argue that all organisms have intrinsic moral status, perhaps in virtue of being teleological centres of life or in some sense possessing a will to live. (A problem for such views flows from the above suggestion that xenobots, while constituted by living cells, are not straightforwardly organisms. However, future biobots may well be.)
Third, some people will claim that creating xenobots (or future biobots) involves “playing God”. Such a claim has multiple possible connotations. It may reduce to the above issues of risk and intrinsic moral status. Alternatively, concerns about “playing God” may be akin to criticisms that engineering life from the “ground up” represents, or promotes, a disrespect for life. This type of criticism has been challenged. Nevertheless, biobots may elicit this response on account of the range of features – from their AI-based design to their “ground up” modular construction – that prompts people to call them “programmable living robots”.
Further reflection on such ethical considerations should inform any future guidelines or regulations for the development and use of biobot technology.