The introduction of non-indigenous plants, animals and pathogens is a pressing global environmental challenge. Although not all introduced species become established and the fraction of those that do often have little appreciable effect on their new ecosystems, many others exert significant ecological, evolutionary and economic impacts. Stimulating further debate, Gozlan [Fish and Fisheries (2008) Vol. 9, pp. 106–115] argued that the majority of intentional freshwater fish introductions associated with aquaculture (fish species providing societal benefits) have not been reported as having an ecological impact. We find little to argue with his suggestion that low risk of ecological impact coupled with high market value encourages further introductions. But do we have an adequate understanding of the ecological risks associated with fish introductions to support such decisions? Indeed, resource managers and decision makers require some scientific knowledge to support their management actions; without this information, a precautionary approach is the only sensible course of action. The precautionary approach implies that the lack of scientific certainty is reason enough for postponing intentional introduction of non-native species to avoid potentially serious or irreversible harm to the environment. Here, we suggest that we actually know very little about ecological impacts associated with fish introductions and that it would be therefore wholly inappropriate to equate a lack of data with a conclusion of ‘no impact’. We discuss four major challenges for enhancing the assessment of risks posed by non-native freshwater fishes in the face of scientific uncertainty and highlight research opportunities and some alternative approaches for confronting these challenges in the future.