Evaluating social media portrayals of imperiled problem species to devise improved conservation messaging strategies

Understanding public sentiment toward wildlife is important for developing effective conservation practices and policies, and securing funding support. This holds true especially in the case of species such as the endangered dhole (Cuon alpinus), which carries a perception debt from centuries of negative attitudes. To examine whether historical negative perceptions are being amplified online, we applied a novel natural language processing (NLP) approach to >6,000 social media posts from Instagram, X (formerly Twitter), Facebook, and YouTube. We compared sentiment and engagement patterns associated with dholes against three other sympatric carnivores that vary in charisma and cultural visibility: Bengal tiger (Panthera tigris), sloth bear (Melursus ursinus), and golden jackal (Canis aureus). We quantified caption descriptors and sentiment across species and platforms, finding that dholes were disproportionately associated with negative adjectives and attitudes. Contrary to expectations, negative captions did not consistently accompany threatening or distressing visuals, highlighting a more complex relationship between language and media. Although audience engagement varied across thematic categories for some species, these differences were relatively small. Visual media outperformed text-based content in eliciting online engagement. Together, these findings demonstrated how digital perceptions of wildlife are shaped by not only species identity but also content framing, platform dynamics, and media format. By integrating scalable NLP tools with targeted media analyses, we developed a replicable framework for decoding online sentiment and provided empirical insights that can inform conservation messaging strategies. Such approaches can be beneficial to researchers and organizations in rebalancing attention toward overlooked species and strengthening conservation narratives.