Related Experiment Video
Updated: Aug 5, 2026
03:14
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
Evaluation of an Artificial Intelligence Language Model for Generating Informed Consent Forms in Oral and
Anuj Jain1,2,3, Rahul Deshmukh4, Umesh Bhutekar5
1Department of Oral and Maxillofacial Surgery, Geetanjali Dental and Research Institute, Geetanjali University, Eklingpura, Udaipur, Rajasthan 313001 India.
Journal of Maxillofacial and Oral Surgery
|July 30, 2026
Summary
AI-generated informed consent forms for oral surgery show good readability but need improvement in ethical and safety aspects. Clinician review is essential before widespread use of these large language model tools.
Area of Science:
- Medical Ethics
- Oral and Maxillofacial Surgery
- Artificial Intelligence
Background:
- Informed consent is crucial in oral and maxillofacial surgery (OMFS), but traditional documents often have poor readability and lack comprehensive ethical/safety details.
- Large language models (LLMs) show potential for drafting consent forms, yet their suitability for OMFS clinical practice requires further investigation.
Purpose of the Study:
- To evaluate the quality of AI-generated informed consent drafts for OMFS procedures using the QUEST framework.
- To assess the readability and clinician acceptance of AI-generated consent forms in OMFS.
Main Methods:
- An in silico study evaluated AI-generated consent drafts for four OMFS procedures (third molar extraction, dental implant, mandibular fracture fixation, soft tissue biopsy).
- Two OMFS surgeons rated drafts using the QUEST framework (Quality, Utility, Ethics, Safety, Transparency) on a 5-point Likert scale.
- Readability was assessed using grade-level metrics, and inter-rater reliability was determined using Cohen's kappa.
Main Results:
- AI-generated drafts had readability scores between grade 6.0-8.0 (mean 6.85).
- Substantial inter-rater reliability was observed (κ=0.78, p<0.001).
- Quality domain scored highest (mean 4.1), while Ethics, Safety, and Transparency domains received lower ratings.
Conclusions:
- AI-generated informed consent drafts offer acceptable readability and procedural clarity for OMFS.
- Limitations exist in ethical, safety, and transparency aspects, necessitating clinician oversight.
- LLMs can aid consent drafting, but further validation is required for clinical implementation.