Adaptive management (AM) emerged in the literature in the mid-1970s in response both to a realization of the extent of uncertainty involved in management, and a frustration with attempts to use modelling to integrate knowledge and make predictions. The term has since become increasingly widely used in scientific articles, policy documents and management plans, but both understanding and application of the concept is mixed. This paper reviews recent literature from conservation and natural resource management journals to assess diversity in how the term is used, highlight ambiguities and consider how the concept might be further assessed. AM is currently being used to describe many different management contexts, scales and locations. Few authors define the term explicitly or describe how it offers a means to improve management outcomes in their specific management context. Many do not adhere to the idea as it was originally conceived, despite citing seminal work. Significant confusion exists over the distinction between active and passive approaches. Over half of the studies reporting to implement AM claimed to have done so successfully, yet none quantified specific benefits, or costs, in relation to possible alternatives. Similarly those studies reporting to assess the approach did so only in relation to specific models and their parameterizations; none assessed the benefits or costs of AM in the field. AM is regarded by some as an effective and well-established framework to support the management of natural resources, yet by others as a concept difficult to realize and fraught with implementation challenges; neither of these observations is wholly accurate. From a scientific and technical perspective many practical questions remain; in particular real-world assessments of the value of experimentation within a management framework, as well as of identified challenges and pathologies, are needed. Further discussion and systematic assessment of the approach is required, together with greater attention to its definition and description, enabling the assessment of new approaches to managing uncertainty, and AM itself.