It is well known that group living holds many advantages, but is collective intelligence one of these? Is a group more intelligent than its smartest individual? This question is not well-defined and, thus, difficult to address. In this project, we approached these questions by studying collective problem-solving by groups of ants.
We studied these questions by considering longhorn crazy ant cooperative transport a behavior in which many ants cooperate to carry a large item of food as a group. This behavior provides a good route for studying collective cognition since, in a natural setting, to get the load to their nest the ants must overcome a large number of highly variable open-ended environmental challenges. An important part of our project was to challenge ant groups with numerous puzzles inspired by different environmental challenges. These allowed us to compare between the performances of different group size (in some cases, down to a single individual) as well as try to decipher the ways in which ants coordinate to overcome these challenges. The main puzzles we studied were a binary decision between two alternative routes; the ant-in-a-labyrinth puzzle: navigation through highly disordered environments; pebble clearing, cooperative transport that requires preparatory modification of the environment, and the piano movers’ puzzle where an odd-shaped “piano” must be maneuvered within a tight environment. To summarize our findings in the briefest manner, ants display collective cognitive skills that make them very general problem-solvers. Importantly, these collective emergent problem-solving capabilities do not erase individual cognition but rather use it and feed back into it to maximally increase performance.
These observations led to and were backed by experiments that were specifically designed to understand how ants achieve increased capabilities. Our experiments strengthened our initial assumptions on the ways ants coordinate their pulling efforts demonstrating that most of the required communication comes from the pulling itself. We further find, that ants actively react to the progress of the load, for example, decreasing their energy use when things are running smoothly. On the group level, we combined a model of individual ants with a physics-engine typically used to simulate motions in computer games. This allowed us to prove that communication by forces is the main source of coordination between ants and understand how the group reacts to constraints. We also constructed a robot that mimics an individual ant and tests her effect on the group. This led to the first experimental evidence ever of the common hypothesis that at criticality, the responsiveness of animal groups to external perturbations is maximized.
Another rare species that performs cooperative transport is our species. We challenged people with the exact same pian-movers-problem that was given to ants. Here again, the puzzle was presented to both individuals and groups of various sizes. We found that, when people are not allowed to communicate freely, group performance deteriorates below that of the average individual person. This happens because, in large groups, people forsake their individual understanding and opinions in favor of achieving consensus. If allowed to speak a group can save its performance to match (but never exceed) that of an individual.
On the theoretical front we have studied ways to measure information flows and the degree of synergy within large coordinated groups. We have developed tools for studying information loops within communication networks and set a much sought after, theoretical basis for the statement that “the whole is larger than the sum of its parts”.
Our results were published in high-end journals and many others are currently under review. Our work has been disseminated in a large number of talks, mainly in Europe and the United States, by the PI and other project members. Our human cooperation experiments were accompanied by a large outreach project in which thousands of participants were exposed firsthand to ant cooperation and collective cognition. Further on the outreach front, some of the ideas inspired by this project ended up in an educational scientific comic book that includes fifty activities with live ants that just about anyone can do and try to understand.
