Evaluation of artificial intelligence tools for the detection of herb–drug interactions

The use of medicinal plants has been increasing both globally and in Türkiye. Accurate identification of medicinal plants is essential for their safe use. Morphologically similar species may possess markedly different active constituents and toxicity profiles. his study aimed to assess the quality of responses generated by ChatGPT 5.5 and Gemini 3 regarding the identification of thirty medicinal plants commonly used in Türkiye and their potential herb–drug interactions. In the botanical evaluation, the mean score of ChatGPT 5.5 was 4.50 ± 0.78 and that of Gemini was 4.30 ± 0.92. The median score was 5 for both systems. In the herb–drug interaction assessment, the mean scores were 4.80 ± 0.41 and 4.00 ± 0.59 for ChatGPT and Gemini 3. ChatGPT 5.5 showed superior performance in the assessment of herb–drug interactions and did not receive a score below 4 for any of the evaluated questions. ChatGPT 5.5 and Gemini 3 demonstrated promising capabilities in identifying medicinal plants commonly used in Türkiye. However, ChatGPT 5.5 provided more accurate and comprehensive information regarding herb–drug interactions.