Collective intelligence: A unifying concept for integrating biology across scales and substrates
Importantly, the definition of intelligence as the ability to reach the same endpoint despite internal or external changes emphasizes not only robustness (successful use of novel navigational policies to overcome perturbations) but also its failure modes. Numerous ways of targeting of its sensory, memory, decision-making, or other components can de-rail the performance of a collective intelligence, resulting in birth defects and malformations. This is quite consistent with the proposed symmetry between the behavioral and developmental domains, because computational neuroscience and cognitive science are replete with interesting ways to think about how cognitive systems make mistakes. The use of tools and concepts across fields has begun, including attempts to understand cancer as a dissociative identity disorder of the morphological collective intelligence109, the use of serotonin reuptake inhibitors and hallucinogens to perturb non-neural development182,183, the modeling of the unstable phenotypes in planarian regeneration as perceptual bistability131, and the finding that some visual illusions that plague vertebrate nervous systems are recapitulated in collective intelligences such as ants184,185. We expect that many concepts from the behavioral sciences that explain failures of learning, recall, Bayesian updating of dynamic signaling models, attention, arousal, and perception will find application in explaining and controlling defects in navigation of anatomical space towards healthy, optimal outcomes.
Another hallmark of collective intelligence is the ability of the higher-level agent to make decisions based on extended patterns of information. For example, in the frog embryo brain, it is the spatial difference in voltage between regions that drives downstream gene expression, not the absolute value of any cells186,187. In other words, cells have to read whole cell fields and recognize specific patterns to determine what to do – the collectivity is seen in the input, as well as the output, of cell groups’ behaviors.
Future work is essential to understand how higher-order entities (organisms, organs, tissues, etc.) distort the energy landscape for their subunits, benefitting from their competencies to navigate spaces of which the subunits are unaware. This underlies the harnessing of cellular signaling and computational abilities to regulative development and regeneration, which implement organ-level homeostatic loops that keep large-scale order against cellular defections (aging and cancer106,188) and injury189. Living matter is a kind of agential material with the ability to propagate information across scales – a phenomenon which has many implications for evolution9, and for bioengineering21.
Many tools are becoming available which increase insight and cross-fertilization of approaches across disciplines. Examples include optogenetic interrogation of single cell78,190–193 and embryonic194–199 dynamics, as well as the very elegant electrotactic ‘SCHEEPDOG’ system which is able to precisely steer collectives of keratinocytes using patterned dynamic electric fields200 that distinguish between collective and individual cell behaviors. In addition to technologies, important additions are conceptual tools, such as the active inference framework201–203 and tools of causal information theory204–212, which will have many applications in the biological sciences.
Future work in this area will also continue to be enriched by advances in the collective intelligence of animal behavior46,213 as well as in the field of swarm robotics214–217. Additional directions for investigation include: how conflict (competition) is used for coordination in collectives6,218, and how propagation of shared stress181,219,220 and the sharing of cellular memories via gap junctions4,109 establish higher-order individuals.
One of the most exciting aspects of this emerging field is the way in which collective intelligence serves as a focal point for exploring the symmetries between developmental biology and neuroscience26. This ranges from the use of cognitive science formalisms to understand morphogenesis and its disorders55,131,221 to the questions of how many human Selves can be sustained by the excitable medium of a human brain67 and the parallels to the multiple bodies that can emerge from a single embryonic blastoderm103.
Many of the same mechanisms (e.g., electrophysiological networks) and control policies are re-used by evolution to bind neurons to collective behavior and animal navigation of 3D space and to bind pre-neural cells to move the body configuration in morphospace5,70. Turing was prescient in studying both intelligence and the chemical basis of self-organization222,223, as the problem of self-organization in familiar neural-based intelligences may have much in common with the problem of self-organizing a non-neural collective intelligence of morphogenesis224. If true, a number of fields can look forward to exciting advances. Cancer, a kind of dissociative identity disorder of the somatic collective intelligence109, limitations in regenerative ability, and many physiological disorders could all be advanced by techniques that exploit not just the low-level mechanisms, but also the higher-level decision-making of life16,17. Neuroscience can benefit from a glimpse into the evolutionary past of the brain’s remarkable capabilities, while developmental biology and bioengineering can borrow the practical and conceptual tools of neuroscience which is likely to be about much more basic principles than the function of classical neurons. Understanding how evolution works in an agential, multiscale material (where it can take advantage of cross-level computation) will nicely complement the efforts of engineers to build and control swarms of robots and AI systems, but who as yet largely work with passive matter where competency exists only at one scale.
Taken together, collective intelligence is an extremely exciting and interdisciplinary emerging field that spans from the most fundamental philosophical problems of the parts-whole relationship to advancing fundamental and applied discovery in a number of important subfields.
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
We thank Douglas Blackiston, Randall Ellis, and Patrick Erickson for their helpful comments on the manuscript, as well as Julia Poirier for editorial assistance. M.L. gratefully acknowledges support of the Guy Foundation Family Trust (103733-00001), of grants 62212 and 62230 from the John Templeton Foundation (the opinions expressed in this publication are those of the author(s) and do not necessarily reflect the views of the John Templeton Foundation), of the Templeton World Charity Foundation (TWCF0606) and of the Air Force Office of Scientific Research under award number FA9550-22-1-0465, Cognitive & Computational Neuroscience program.
M.L. and P.M. wrote this Perspective together, including working on the text and creating the figures.
Communications Biology thanks Timothy Jackson and Ricard Solé for their contribution to the peer review of this work. Primary Handling Editor: Manuel Breuer. A peer review file is available.
The authors declare the following competing interests: Tufts University has a sponsored research agreement with a company, Astonishing Labs, to fund projects relevant to the collective intelligence of cells.
The online version contains supplementary material available at 10.1038/s42003-024-06037-4.