Evaluation of validity, reliability, and readability of AI chatbots for gestational diabetes mellitus: a multi-model comparative study

Contemporary AI chatbots can generate generally accurate and moderately reliable GDM-related information, with newer model generations showing clear gains in diagnostic validity. However, limited transparency and systematically high reading levels indicate that these tools are not yet suitable as stand-alone resources for GDM patient education and should be used as adjuncts to clinician counseling and professionally curated materials.