Scientific Creativity of Artificial Intelligence: Evaluation of Novel Research Ideas in Oral and Maxillofacial Surgery

Background: Artificial intelligence (AI) has the potential to contribute to scientific research as well as to many other fields. Although the contribution of AI in presenting and analyzing data has been widely studied, its potential to generate research ideas, especially in specialized fields such as oral and maxillofacial surgery, has not been widely studied. This study aims to evaluate the ability of ScholarGPT, an AI model specially trained for academic purposes, to suggest original and unpublished research ideas in six different fields of oral and maxillofacial surgery. Materials and Methods: ScholarGPT was asked to generate five research ideas for each of the following six areas: impacted third molar surgery, dental implant surgery, bone augmentation techniques, orthognathic surgery, distraction osteogenesis, and zygomatic implant surgery. The novelty of these ideas was assessed through a literature search of PubMed. Results: Among the AI-generated research ideas, 53.3% were identified as original. The novelty rate varied by topic, with the highest originality observed in orthognathic surgery and distraction osteogenesis (80%), while the lowest was found in impacted third molar surgery (20%). Statistical analysis revealed no significant correlation between research topics and originality (p = 0.44). Conclusion: ScholarGPT can help generate new research ideas in oral and maxillofacial surgery using existing knowledge, but human evaluation is needed to determine whether these proposals are truly original and applicable.