Generative artificial intelligence (GenAI), including large language models (LLMs), has vast potential applications in health care and education. However, it is unclear how proficient LLMs are in interpreting written input and providing accurate answers in dentistry. This study aims to investigate the accuracy of GenAI in answering questions from dental licensing examinations. A total of 1461 multiple-choice questions from question books for the US and the UK dental licensing examinations were input into 2 versions of ChatGPT 3.5 and 4.0. The passing rates of the US and UK dental examinations were 75.0% and 50.0%, respectively. The performance of the 2 versions of GenAI in individual examinations and dental subjects was analysed and compared. ChatGPT 3.5 correctly answered 68.3% (n = 509) and 43.3% (n = 296) of questions from the US and UK dental licensing examinations, respectively. The scores for ChatGPT 4.0 were 80.7% (n = 601) and 62.7% (n = 429), respectively. ChatGPT 4.0 passed both written dental licensing examinations, whilst ChatGPT 3.5 failed. ChatGPT 4.0 answered 327 more questions correctly and 102 incorrectly compared to ChatGPT 3.5 when comparing the 2 versions. The newer version of GenAI has shown good proficiency in answering multiple-choice questions from dental licensing examinations. Whilst the more recent version of GenAI generally performed better, this observation may not hold true in all scenarios, and further improvements are necessary. The use of GenAI in dentistry will have significant implications for dentist-patient communication and the training of dental professionals.

Gingivitis is one of the most prevalent plaque-initiated dental diseases globally. It is challenging to maintain satisfactory plaque control without continuous professional advice. Artificial intelligence may be used to provide automated visual plaque control advice based on intraoral photographs. Frontal view intraoral photographs fulfilling selection criteria were collected. Along the gingival margin, the gingival conditions of individual sites were labelled as healthy, diseased, or questionable. Photographs were randomly assigned as training or validation datasets. Training datasets were input into a novel artificial intelligence system and its accuracy in detection of gingivitis including sensitivity, specificity, and mean intersection-over-union were analysed using validation dataset. The accuracy was reported according to STARD-2015 statement. A total of 567 intraoral photographs were collected and labelled, of which 80% were used for training and 20% for validation. Regarding training datasets, there were total 113,745,208 pixels with 9,270,413; 5,711,027; and 4,596,612 pixels were labelled as healthy, diseased, and questionable respectively. Regarding validation datasets, there were 28,319,607 pixels with 1,732,031; 1,866,104; and 1,116,493 pixels were labelled as healthy, diseased, and questionable, respectively. AI correctly predicted 1,114,623 healthy and 1,183,718 diseased pixels with sensitivity of 0.92 and specificity of 0.94. The mean intersection-over-union of the system was 0.60 and above the commonly accepted threshold of 0.50. Artificial intelligence could identify specific sites with and without gingival inflammation, with high sensitivity and high specificity that are on par with visual examination by human dentist. This system may be used for monitoring of the effectiveness of patients' plaque control.

Computer-aided design and computer-aided manufacturing (CAD-CAM) technology has greatly improved the efficiency of the fabrication of dental prostheses. However, the design process (CAD stage) is still time-consuming and labor intensive. The purpose of this feasibility study was to investigate the accuracy of a novel artificial intelligence (AI) system in designing biomimetic single-molar dental prostheses by comparing and matching them to the natural molar teeth. A total of 169 maxillary casts were obtained from healthy dentate participants. The casts were digitized, duplicated, and processed with the removal of the maxillary right first molar. A total of 159 pairs of original and processed casts were input into the Generative Adversarial Networks (GANs) for training. In validation, 10 sets of processed casts were input into the AI system, and 10 AI-designed teeth were generated through backpropagation. Individual AI-designed teeth were then superimposed onto each of the 10 original teeth, and the morphological differences in mean Hausdorff distance were measured. True reconstruction was defined as correct matching between the AI-designed and original teeth with the smallest mean Hausdorff distance. The ratio of true reconstruction was calculated as the Intersection-over-Union. The reconstruction performance of the AI system was determined by the Hausdorff distance and Intersection-over-Union. Data of validation showed that the mean Hausdorff distance ranged from 0.441 to 0.752 mm and the Intersection-over-Union of the system was 0.600 (60%). This study demonstrated the feasibility of AI in designing single-molar dental prostheses. With further training and optimization of algorithms, the accuracy of biomimetic AI-designed dental prostheses could be further enhanced.

Objectives To evaluate the long-term longevity and patient-reported outcomes of two-unit cantilevered (CL2) and three-unit fixed–fixed (FF3) resin-bonded fixed partial dentures (RBFPDs) for the replacement of a maxillary permanent incisor. Materials and methods Twenty-eight subjects were randomly assigned to receive either a CL2 or FF3 RBFPD placed by one operator. Prosthesis longevity was determined by clinical examination and history. Success was defined as absence of complications requiring intervention and survival as retention of the original prosthesis in mouth. Subjects’ satisfaction was assessed using visual analogue scale (VAS) and oral health-related quality of life (OHRQoL) using Oral Health Impact Profile (OHIP-49). Outcomes were analysed with t-test/Mann–Whitney U test, chi-square and log-rank test at significance level α = 0.05. Results Twenty-two subjects were reviewed. Thirteen of fifteen CL2 and ten of fourteen FF3 RBFPDs were examined (79.3 percent response rate) with a mean service life of 216.5 ± 20.8 months. All CL2 RBFPDs survived with no complications while only 10 percent of FF3 experienced no complications and only 50 percent of them survived (both P = 0.000). CL2 had a significantly better success and survival rate than FF3 (P = 0.000 and P = 0.009, respectively). There was no significant difference in subjects’ satisfaction and OHRQoL apart from CL2 group subjects had a higher satisfaction in cleaning of the prosthesis (84.1 ± 13.6) than FF3 group (72.6 ± 11.7) (P = 0.05). Conclusions Two-unit cantilevered RBFPDs were observed to have a significantly better success and survival than the FF3 design for the replacement of a maxillary incisor. Good patient-reported outcomes have been found for RBFPDs in single-tooth replacement in aesthetic zone.