This study aimed to develop a generative AI-based intelligent counseling chatbot system for doping prevention and to evaluate its performance. The retrieval corpus was constructed from publicly available resources such as regulations and prohibited substance lists available on the Korea Anti-Doping Agency (KADA) website. The system was constructed by combining the Qwen3:30B-A3B model with a vector database (ChromaDB). For performance evaluation, 51 test items were extracted from KADA's Frequently Asked Questions (FAQ) and “Voice of Customers” sections. The analysis results showed that the proposed system achieved an overall accuracy of 82.35% and a response consistency of 96.08%, demonstrating high performance. This system can provide counseling and educational services to diverse groups—including athletes, coaches, and parents—regardless of time and location, thereby enhancing the efficiency and accessibility of doping prevention administration.
SaragihZKJunaidiSamsudin. Enhancing athletes’ anti-doping education through mobile applications: a review of current strategies and emerging technologies. KnE Soc Sci2024; 9: 680–684.
2.
LeeSYParkJHYoonJ, et al.A validation study of a deep learning-based doping drug text recognition system to ensure safe drug use among athletes. Healthcare2023; 11: 1769.
3.
LeeJKangSAnD, et al.A meta-analysis of attitudes toward doping among Korean athletes. The Asian Journal of Kinesiology2024; 26: 86–95.
4.
NichollsARLazurasLPetrouM, et al.A systematic review on the effectiveness of anti-doping education for university students. Emerging Trends in Drugs, Addictions, and Health2025; 5: 100168.
5.
YunHJParkJHYoonJ, et al.Current Practices of Herbal Medicine Prescription for Athletes and Doping Awareness Among Korean Medicine Doctors. J Drug Issues2025. 10.1177/0022042625136437
6.
TsivitanidouOChristodoulidesEPetrouM. High-School Athletes’ knowledge, attitudes, and perceptions on doping: the Cyprus sport-school study. Youth2023; 3: 596–622.
7.
PöppelKBüschD. Every young athlete counts: are tailored doping prevention programs necessary in young elite sports?Front Sports Act Living2022; 4: 858730.
8.
LeeJYParkJHKimC. Development of an AI-Based Model for Identifying Doping Drugs Through Pill Image Analysis. J Drug Issues2025. 10.1177/00220426251352781
9.
ParkJHYoonSHYoonJ, et al.Development of doping drug recognition system : application of deep learning-based OCR technology. The Korean Journal of Physical Education2022; 61: 83–92.
10.
AkpanIJKobaraYMOwolabiJ, et al.Conversational and generative artificial intelligence and human–chatbot interaction in education and research. Int Trans Oper Res2025; 32: 1251–1281.
11.
ChagasBAPaganoASPratesRO, et al.Evaluating user experience with a chatbot designed as a public health response to the COVID-19 pandemic in Brazil: mixed methods study. JMIR Hum Factors2023; 10: e43135.
12.
ChenSGuevaraMMoningiS, et al.The effect of using a large language model to respond to patient messages. Lancet Digit Health2024; 6: e379–e381.
13.
AbbasianMKhatibiEAzimiI, et al.Foundation metrics for evaluating effectiveness of healthcare conversations powered by generative AI. NPJ Digit Med2024; 7: 82.
14.
BackhouseSHGriffithsCMcKennaJ. Tackling doping in sport: a call to take action on the dopogenic environment. Br J Sports Med2018; 52: 1485–1486.
15.
PanMXiongWZhouS, et al.LLM-Based Query expansion with Gaussian kernel semantic enhancement for dense retrieval. Electronics (Basel)2025; 14: 1744.
16.
WeiJWangXSchuurmansD, et al.Chain-of-thought prompting elicits reasoning in large language models. Adv Neural Inf Process Syst2022; 35: 24824–24837.
17.
JiZLeeNFrieskeR, et al.Survey of hallucination in natural language generation. ACM Comput Surv2023; 55: 1–38.
18.
KarpukhinVOguzBMinS, et al. Dense passage retrieval for open-domain question answering. In: Proceedings of the 2020 conference on empirical methods in natural language processing (EMNLP), 2020; pp.6769–6781. Online: Association for Computational Linguistics.
19.
KimJHChoiYS. Lightweight Pre-trained Korean language model based on knowledge distillation and low-rank factorization. Entropy2025; 27: 79.
20.
LewisPPerezEPiktusA, et al.Retrieval-augmented generation for knowledge-intensive nlp tasks. Adv Neural Inf Process Syst2020; 33: 9459–9474.
21.
YangALiAYangB, et al. Qwen3 technical report. Report, arXiv, 2025, arXiv:2505.09388.10.48550/arXiv.2505.09388
22.
ZhuSHuWYangZ, et al.Qwen-2.5 outperforms other large language models in the Chinese national nursing licensing examination: retrospective cross-sectional comparative study. JMIR Med Inform2025; 13: e63731.
23.
BeckSKuhnerMHaarM, et al.Evaluating the accuracy and reliability of AI chatbots in disseminating the content of current resuscitation guidelines: a comparative analysis between the ERC 2021 guidelines and both ChatGPTs 3.5 and 4. Scand J Trauma Resusc Emerg Med2024; 32: 95.
24.
ÖzcivelekTÖzcanB. Comparative evaluation of responses from DeepSeek-R1, ChatGPT-o1, ChatGPT-4, and dental GPT chatbots to patient inquiries about dental and maxillofacial prostheses. BMC Oral Health2025; 25: 71.
25.
QiuJZhouY. Assessing the accuracy and consistency of answers by ChatGPT to questions regarding carbon monoxide poisoning. PloS One2024; 19: e0311937.
26.
NaziZAPengW. Large language models in healthcare and medical domain: a review. Informatics2024; 11: 57.
27.
ScaffSPReisFJFerreiraGE, et al.Assessing the performance of AI chatbots in answering patients’ common questions about low back pain. Ann Rheum Dis2025; 84: 143–149.
28.
GoodmanRSPatrinelyJRStoneCA, et al.Accuracy and reliability of chatbot responses to physician questions. JAMA Netw Open2023; 6: e2336483–e2336483.
