Search for a command to run...
Abstract Biodiversity assessment using passive acoustic monitoring has historically been challenging due to the limited availability of multi‐species acoustic detectors. In this context, acoustic indices were introduced as an alternative way to represent species diversity in acoustic datasets. Although there is increasing evidence that their effectiveness as a proxy for species diversity is limited, acoustic indices are still used and recommended for biodiversity research. Acoustic indices face structural issues, with a flawed theoretical basis and weak empirical support. Additionally, limited generalizability, dependence on context and inconsistent performance make it difficult to assess their value in the inferential process. Without a clear method to map acoustic indices into meaningful biological components, their interpretability is greatly compromised. Biodiversity assessment through acoustic monitoring benefits significantly from frameworks that use species as the fundamental analytical unit, even when misclassification might occur. In contrast, acoustic indices ignore species recognition due to their reliance on properties such as spectral distribution of acoustic energy and consequently limit the inference of causal links to species diversity. A variety of options are available to enhance species‐based biodiversity acoustic monitoring. We emphasize that manual annotation and single‐species detectors have contributed significantly to the field, and effective methods for multi‐species classification now exist. This has opened new paradigms and expanded opportunities for directly identifying target signals in audio data across a broad range of species. In the ongoing biodiversity crisis, resources should be prioritized for tools and approaches that provide managers and stakeholders with meaningful information on the composition, structure, and function of biodiversity. Acoustic indices are not suitable as reliable biodiversity indicators, and their continued use in biodiversity research is unjustifiable. Acoustic monitoring should leverage existing and emerging methods to extract species‐level data from recordings, offering more valuable and impactful contributions to ecology and conservation efforts.