Scientists are not immune from bias. Studying nonhuman species objectively is inherently challenging, especially for ‘charismatic’ and ostensibly human-like behaviours. Animal tool use is a prime example: while often considered a hallmark of intelligence, the amount of research attention and public interest it generates seems disproportionate when compared with other behaviours involving similar manipulative skills, particularly nest building. Here, we reveal striking disparities in the treatment of tool use and nest building in the animal behaviour literature. We find that tool use publications are more highly cited, are more likely to be published in higher-impact journals and use more terminology suggestive of ‘intelligence’ and human-like cognition compared with nest building publications. Our findings are not confounded by taxonomic biases: these disparities persist even within studies of great apes and Corvus species. Further, we find that articles with more frequent use of ‘intelligent’ terminology are more highly cited, suggesting incentives for the use of anthropomorphic language in scientific articles. Finally, we find that tool use papers are more highly cited than nest building papers even when controlling for the use of ‘intelligent’ language, showing that both language use and behaviour have additive effects on research attention. We argue that these research disparities are partly driven by a widespread assumption that tool use requires more complex cognition than nest building. Since the cognitive mechanisms underpinning either behaviour are still not well understood, we suggest that the widespread appeal of animal tool use is partly due to anthropocentrism.