Comparative performance of neurosurgery-specific, peer-reviewed versus general AI chatbots in bilingual board examinations: evaluating accuracy, consistency, and error minimization strategies
Our findings suggest that GPT-4-based models and AtlasGPT can handle specialized neurosurgery questions at a near-expert level for SBA-R, SBA-I, and True/False formats. Nevertheless, all models exhibit notable limitations in questions with images, indicating that these tools remain supplementary rather than definitive solutions for neurosurgical training and decision-making.
Authors
- Gökberk Erol (ORCID: https://orcid.org/0000-0001-6651-5486)
- Abuzer Güngör (ORCID: https://orcid.org/0000-0002-2792-7610)
- Yücel Doğruel (ORCID: https://orcid.org/0000-0003-4314-5579)
- Mahmut Çamlar (ORCID: https://orcid.org/0000-0002-6467-0172)
- Umut Tan Sevgi (ORCID: https://orcid.org/0000-0003-3875-859X)
- Furkan Karakaş
Institutions
- Ministry of Health (TR)
- Izmir University (TR)
- Dokuz Eylül University (TR)
- Istinye University (TR)
- Education Training And Research (US)
- Sağlık Bilimleri Üniversitesi (TR)
Publication Details
- Journal
- Acta Neurochirurgica
- Published
- 2025-09-09
- DOI
- https://doi.org/10.1007/s00701-025-06628-y
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00