Akademik Veri Yönetim Sistemi
Nordic Journal of Botany, 2026 (SCI-Expanded, Scopus)
The advancement of biological sciences has long been closely linked to technological progress. ChatGPT, a generative artificial intelligence chatbot capable of producing human-like conversational responses, has recently attracted attention as a potential support tool for scientific research. However, its applicability and reliability in natural sciences, particularly in botanical research, remain insufficiently evaluated. In this study, we assess the performance of ChatGPT across several key tasks relevant to botany, including plant identification, construction of taxonomic keys, analysis of morphological descriptions, distribution mapping, and image-based species recognition. Systematic testing conducted across multiple geographic regions, including Asia, Australia, Europe, and North and South America, revealed substantial variability in the quality and accuracy of ChatGPT's outputs. Although the chatbot occasionally generated useful content, such as general identification keys and improved manuscript text, it frequently produced inaccurate or misleading information. Common issues included incorrect species identifications, inconsistent handling of synonymy, fabrication of literature sources, and errors in interpreting taxonomic and distributional data. Performance was particularly limited for closely related taxa, hybrids, and microphotographs, whereas genus-level identification and editorial tasks proved comparatively more reliable. Attempts to generate functional distribution maps or executable code were largely unsuccessful. Overall, ChatGPT cannot currently be considered a reliable tool for independent botanical research. Its limitations pose particular risks for students and early-career researchers who may lack the expertise required to critically evaluate its outputs. Nevertheless, the rapid development of artificial intelligence technologies suggests that future iterations may offer more consistent and accurate support for biodiversity studies. At present, the most appropriate use of ChatGPT in botanical research lies in text editing and general linguistic assistance, provided that all factual information is rigorously verified by experts.