Some Insects Are More Equal Than Others: A Comparison of Popular Large Language Model Chatbots' Treatment of Different Insect Groups

ABSTRACT Bees and butterflies are generally viewed positively in society, whereas other insects, such as wasps or flies, are often underappreciated despite their essential ecological roles. As biodiversity loss continues to pose a major challenge for humankind, it necessitates a re‐evaluation of these biases to support the protection of biodiversity as a whole. With large language model (LLM) chatbots becoming increasingly integrated into daily life for information dissemination and education purposes, understanding their inherent biases is vital. In this study, we tested 10 popular Western LLM chatbots using simple prompts to assess how they portray bees and wasps. Butterflies, moths, flies, and mosquitoes were also included for comparison to evaluate broader societal perceptions. Our results show that bees and butterflies are indeed depicted positively by the LLM chatbots and moths somewhat positively, while wasps, flies, and mosquitoes are associated with more negative portrayals. We found that LLMs mirror prevailing human biases toward different insect groups and their perceived importance in nature conservation. Moreover, we demonstrate that LLM chatbots tend to oversimplify insect diversity by predominantly restricting “bees” to honeybees and “wasps” to yellowjackets, thereby neglecting the broader biodiversity that includes wild bees and parasitoid wasps. The chatbots also appeared to favor Nearctic species when recommending conservation priorities. By highlighting these biases and discussing their implications, our research underscores the importance of nuanced science communication and expert involvement in decision‐making for nature conservation. Addressing such biases is essential to prevent the reinforcement of public misconceptions and to promote the protection of ecologically indispensable yet less popular insect groups.