Collective behavior of animals is a fascinating example of self-organization in biology. In contrast to non-living systems like magnets, where the emergent macroscopic properties of the system are simply a consequence of physical interac:ons between individual components, collective animal behaviors and the underlying social interactions are the result of evolutionary adaptation. Thus, it is believed that collective behavior confers fitness benefits to individuals, for example by promoting exchange of social information, accurate collective decisions, or protection from predators. In this context, it has been argued that animal collectives should operate in a special parameter region close to so-called critical points, i.e. close to transitions between different collective states,
where various aspects of collective information processing become optimal. Here, we will investigate the "criticality hypothesis" by combining model simulations together with laboratory and field experiments on collective predator response in fish.