Erken Çocukluk Çağı Çürüklerinde Yapay Zekâ: Büyük Dil Modeli Tabanlı Sohbet Robotlarının Doğruluk ve Zamansal Tutarlılık Performansı

Abstract

Amaç: Bu çalışmanın amacı, büyük dil modeli (LLM) tabanlı yeni nesil üretken yapay zekâ (AI) araçlarının erken çocukluk çağı çürükleri (ECC) ile ilgili verdiği yanıtların doğruluğunu değerlendirmek ve zamansal tutarlılık açısından performanslarını incelemektir. Yöntem: Amerikan Pediatrik Diş Hekimliği Akademisi (AAPD) Politikası doğrultusunda geliştirilen ECC ile ilgili 21 adet doğru/yanlış sorusu, 3 gün boyunca hem halka açık olan hem de ücretli üyelik gerektiren beş sohbet robotuna (ChatGPT sürüm 5.2 (Open AI), DeepSeek (DeepSeek AI), Perplexity (Perplexity AI), Gemini (Google) ve Copilot (Bing/Microsoft)) yöneltilmiştir. Yanıtlar kaydedilmiş ve doğru cevaplarla karşılaştırılmıştır. Bulgular: Tüm sohbet robotları benzer doğruluk performansı göstermiştir (p>0,05). Sohbet robotlarının performansı, gününe veya saatine göre değişmektedir (p<0,05). Performansın en düşük olduğu zamanlar, 2.gün sabah ve 3.gün sabahtır (p<0,05). Tüm zaman dilimi boyunca en yüksek stabilite ChatGPT-5.2’de, en düşük stabilite ise Perplexity’de gözlenmiştir. Sonuç: Bu çalışma, kanıta dayalı çocuk diş hekimliğini desteklemede LLM’lerin büyük potansiyele sahip olduğunu gösterse de; AI araçlarının kesin bilimsel bilgi yerine geçecek şekilde değil, ihtiyaç halinde destekleyici nitelikte araçlar olarak dikkatli bir şekilde kullanılması gerektiğini vurgulamaktadır.

Keywords

• Büyük dil modelleri
• Yapay zeka
• Chatbot
• Erken çocukluk çağı çürüğü
• Çocuk diş hekimliği
• Sağlık okuryazarlığı

Artificial Intelligence in Early Childhood Caries: Accuracy and Temporal Consistency Performance of Large Language Model–Based Chatbots

Abstract

Objective: To evaluate the accuracy of responses provided by large language model (LLM)-based next-generation generative artificial intelligence (AI) tools regarding early childhood caries (ECC) and to examine their performance in terms of temporal consistency. Method: Twenty-one true/false questions on ECC, developed in accordance with the policy of the American Academy of Pediatric Dentistry (AAPD), were posed over three days to five chatbot systems, including both publicly accessible and subscription-based platforms: ChatGPT version 5.2 (OpenAI), DeepSeek (DeepSeek AI), Perplexity (Perplexity AI), Gemini (Google), and Copilot (Bing/Microsoft). The responses were recorded and compared with the correct answers. Results: All chatbots demonstrated similar accuracy performance (p>0.05). Chatbots’ performance varied according to the day and time (p<0.05). The lowest performance was observed on the morning of Day 2 and the morning of Day 3 (p<0.05). Across all time points, the highest stability was observed in ChatGPT-5.2, whereas the lowest stability was observed in Perplexity. Conclusion: Although this study demonstrates that LLMs have considerable potential to support evidence-based pediatric dentistry, it emphasizes that AI tools should not be used as substitutes for definitive scientific knowledge, but rather should be employed cautiously as supportive tools when needed.

Keywords

• Large language models
• Artificial intelligence
• Chatbot
• Early childhood caries
• Pediatric dentistry
• Health literacy

References

1. 1- American Academy of Pediatric Dentistry. Policy on early childhood caries (ECC): Consequences and preventive strategies. The Reference Manual of Pediatric Dentistry. Chicago, IL: American Academy of Pediatric Dentistry; 2025:96-100.
2. 2- American Academy of Pediatric Dentistry. Policy on early childhood caries (ECC): Unique challenges and management considerations. The Reference Manual of Pediatric Dentistry. Chicago, IL: American Academy of Pediatric Dentistry; 2025:101-3.
3. 3- Park YH, Kim SH, Choi YY. Prediction models of early childhood caries based on machine learning algorithms. Int J Environ Res Public Health 2021;18(16):8613. https://doi.org/10.3390/ijerph18168613
4. 4- Zheng J, et al. Unlocking the potentials of large language models in orthodontics: a scoping review. Bioeng 2024;11(11):1145. https://doi.org/10.3390/bioengineering11111145
5. 5- Claman D, Sezgin E. Artificial intelligence in dental education: opportunities and challenges of large language models and multimodal foundation models. JMIR Med Educ 2024;10(1):e52346. https://doi.org/10.2196/52346
6. 6- Dermata A, et al. Evaluating the evidence-based potential of six large language models in paediatric dentistry: a comparative study on generative artificial intelligence. Eur Arch Paediatr Dent 2025;26(3):527-35. https://doi.org/10.1007/s40368-025-01012-x
7. 7- Balel Y. Can ChatGPT be used in oral and maxillofacial surgery?. J Stomatol Oral Maxillofac Surg 2023;124(5):101471. https://doi.org/10.1016/j.jormas.2023.101471
8. 8- Goodman RS, et al. Accuracy and reliability of chatbot responses to physician questions. JAMA Netw Open 2023;6(10):e2336483. doi:10.1001/jamanetworkopen.2023.36483

9. 9- Abuabara A, et al. Evaluating the accuracy of generative artificial intelligence models in dental age estimation based on the Demirjian's method. Front Dent Med 2025;6:1634006. https://doi.org/10.3389/fdmed.2025.1634006
10. 10- Gravina AG, et al. Charting new AI education in gastroenterology: cross-sectional evaluation of ChatGPT and perplexity AI in medical residency exam. Dig Liver Dis 2024;56(8):1304-11. https://doi.org/10.1016/j.dld.2024.02.019
11. 11- Zaorska K, et al. Prediction of early childhood caries based on single nucleotide polymorphisms using neural networks. Genes 2021;12(4):462. https://doi.org/10.3390/genes12040462
12. 12- Koopaie M, et al. Salivary cystatin S levels in children with early childhood caries in comparison with caries-free children; statistical analysis and machine learning. BMC Oral Health 2021;21(1):650. https://doi.org/10.1186/s12903-021-02016-x
13. 13- Kapoor D, Garg D, Tadakamadla SK. Brush, byte, and bot: quality comparison of artificial intelligence-generated pediatric dental advice across ChatGPT, Gemini, and Copilot. Front Oral Health 2025;6:1652422. https://doi.org/10.3389/froh.2025.1652422
14. 14- Kusaka S, et al. Usefulness of generative artificial intelligence (AI) tools in pediatric dentistry. Diagnostics 2024;14(24):2818. https://doi.org/10.3390/diagnostics14242818
15. 15- Mukhopadhyay A, Mukhopadhyay S, Biswas R. Evaluation of large language models in pediatric dentistry: A Bloom’s taxonomy-based analysis. Folia Medica 2025;67(4):e154338. https://doi.org/10.3897/folmed.67.e154338
16. 16- American Academy of Pediatric Dentistry. Oral health policies & recommendations. (2025-2026). Received from: https://www.aapd.org/research/oral-health-policies--recommendations/
17. 17- Suárez A, et al. Beyond the Scalpel: Assessing ChatGPT's potential as an auxiliary intelligent virtual assistant in oral surgery. Comput Struct Biotech J 2024;24:46-52. https://doi.org/10.1016/j.csbj.2023.11.058
18. 18- Suárez A, et al. Unveiling the ChatGPT phenomenon: evaluating the consistency and accuracy of endodontic question answers. Int Endod J 2024;57(1):108-13. https://doi.org/10.1111/iej.13985
19. 19- Ozden I, et al. Assessment of artificial intelligence applications in responding to dental trauma. Dent Traumatol 2024;40(6):722-29. https://doi.org/10.1111/edt.12965
20. 20- Akitomo T, et al. Generative Artificial Intelligence and Large Language Models in Paediatric Dentistry: A Scoping Review. Int Dent J 2026;76(2):109401. https://doi.org/10.1016/j.identj.2025.109401
21. 21- Tinanoff N, et al. Early childhood caries epidemiology, aetiology, risk assessment, societal burden, management, education, and policy: Global perspective. Int J Paediatr Dent 2019;29(3):238-48. https://doi.org/10.1111/ipd.12484
22. 22- Pang L, et al. A new model for caries risk prediction in teenagers using a machine learning algorithm based on environmental and genetic factors. Front Genet 2021;12:636867. https://doi.org/10.3389/fgene.2021.636867
23. 23- Gökcek Taraç M, Nale T. Artificial intelligence in pediatric dental trauma: do artificial intelligence chatbots address parental concerns effectively?. BMC Oral Health 2025;25(1):736. https://doi.org/10.1186/s12903-025-06105-z
24. 24- Ghahramani Z. Probabilistic machine learning and artificial intelligence. Nature 2015;521(7553):452-59. https://doi.org/10.1038/nature14541
25. 25- Lisboa RM, et al. Performance of five free large language models in dental trauma: a 30-day longitudinal benchmark study. Front Oral Health 2025;6:1737114. https://doi.org/10.3389/froh.2025.1737114
26. 26- Mohammad‐Rahimi H, et al. Validity and reliability of artificial intelligence chatbots as public sources of information on endodontics. Int Endod J 2024;57(3):305-14. https://doi.org/10.1111/iej.14014
27. 27- Sismanoglu S, Capan BS. Performance of artificial intelligence on Turkish dental specialization exam: can ChatGPT-4.0 and gemini advanced achieve comparable results to humans?. BMC Med Educ 2025;25(1):214. https://doi.org/10.1186/s12909-024-06389-9
28. 28- Karamüftüoğlu N, Varol EA, Bal C. Exploring artificial intelligence chatbots in pediatric fluoride education: a cross-sectional study. Sci Rep 2025;16(1):182. https://doi.org/10.1038/s41598-025-28857-y
29. 29- Isik V, Sengul RI, Sismanoglu S. Evaluation of Large Language Model‐Based Chatbots for Dental Trauma Management: A Comparative Study Based on Accuracy, Consistency and Information Quality. Aust Endod J 2026;52(2):356-63. https://doi.org/10.1111/aej.70055
30. 30- Sezer B, Aydoğdu T. Performance of advanced artificial intelligence models in pulp therapy for immature permanent teeth: A comparison of ChatGPT-4 Omni, DeepSeek, and Gemini Advanced in accuracy, completeness, response time, and readability. J Endod 2025;51(11):1675-84. https://doi.org/10.1016/j.joen.2025.08.011
31. 31- Bashah A, et al. Evaluation of deepseek, gemini, ChatGPT-4o, and perplexity in responding to salivary gland cancer. BMC Oral Health 2025;25(1):1358. https://doi.org/10.1186/s12903-025-06726-4
32. 32- Mine Y, et al. Beyond exam accuracy: Tracking a persistent-failure set reveals visual dental reasoning gaps in multimodal LLMs. J Dent 2026;170:106675. https://doi.org/10.1016/j.jdent.2026.106675