Human activities have altered the composition of species assemblages through the introduction of non-native species and the extinction of threatened species. However, it remains unclear whether non-native species can compensate for the loss of threatened species and thus maintain ecosystem functioning. Here we tested whether non-natives are functionally and/or phylogenetically similar (compensation hypothesis) or distinct (shift hypothesis) from native and threatened species on bird assemblages in 267 regions worldwide. We show that non-native species were more functionally distinct from threatened species than expected by chance but more phylogenetically related. Globally, this results in an increase in the functional richness of bird assemblages due to the introduction of new functional traits but a decrease in the phylogenetic richness due to the potential loss of phylogenetically unique threatened species. Furthermore, these patterns vary across continents, revealing the role of human history and footprint across the world and outlining priority areas where international bird conservation should focus. In the context of the biodiversity crisis and the increasing number of non-native species worldwide, the changes in the functional and phylogenetic structure of the bird communities might increase the vulnerability of ecosystems.