The dynamics of large ecological systems result from vast numbers of interactions between individual organisms. Here, we develop mathematical theory to show that the rate of such interactions is inherently limited by the ability of organisms to gain information about one another. This phenomenon, which we call 'information limitation', is likely to be widespread in real ecological systems and can dictate both the rates of ecological interactions and long-run dynamics of interacting populations. We show how information limitation leads to sigmoid interaction rate functions that can stabilize antagonistic interactions and destabilize mutualistic ones; as a species or type becomes rare, information on its whereabouts also becomes rare, weakening coupling with consumers, pathogens and mutualists. This can facilitate persistence of consumer-resource systems, alter the course of pathogen infections within a host and enhance the rates of oceanic productivity and carbon export. Our findings may shed light on phenomena in many living systems where information drives interactions.