Scholarlink

Background <p>ChatGPT, an artificial intelligence (AI) chatbot developed by OpenAI, is increasingly being used in healthcare, including dentistry, for patient education; this study aimed to assess the usability and quality of ChatGPT’s responses to pregnancy-related oral health queries in Saudi Arabia.</p> Method <p>This two-part cross-sectional study assessed pregnant Arabic women’s perceptions of ChatGPT for oral health queries and evaluated its responses using an online questionnaire. Responses from ChatGPT-4o mini were rated by 5 dental experts with regard to accuracy, clarity, relevance, and acceptance using a 5-point Likert scale.</p> Results <p>Among the 300 participants, 42.0% (126) knew about ChatGPT, 33.7% (101) had previously used it, 14.3% (43) had used it to obtain medical information, 8.7% (26) had used it for dental information, and 8.3% (25) had used it for dental information during pregnancy. Attitudes regarding ChatGPT were rated from 1 to 4. Except for 1 item, the means were all above the midpoint. Attitude ratings ranged from a mean of 2.71 (SD 0.76) for ChatGPT competency to a mean of 2.34 (SD 0.92) for its ability to replace human interactions. However, ChatGPT competency (<i>P</i> = .028), security (<i>P</i> = .015), willingness to use ChatGPT for inquiries (<i>P</i> = .021), ability to assist in informed decision-making (<i>P</i> = .01), willingness to make decisions based on recommendations (<i>P</i> = .024), and persuasiveness (<i>P</i> = .049) were significantly different based on educational level. Pregnant women with higher levels of education rated these aspects significantly lower than those with a high school diploma or bachelor’s degree.</p> Conclusion <p>ChatGPT provided useful oral health information for pregnant individuals, but its responses required revision and supervision by health professionals. Its usage among pregnant women in Saudi Arabia remained low.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Can ChatGPT be trusted? Evaluating AI responses to oral health questions among pregnant Arabic-speaking women

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月10日

查看全文
  • Khalid Talal Aboalshamat,
  • Jomanh Humied Alnafei,
  • Lojain Ahmed Alkhattabi,
  • Ghadi Yaqoub Alhawsawi,
  • Shrooq Majed Alahmadi,
  • Shatha Omar Almalki,
  • Afnan Anas Nassar

摘要

Background

ChatGPT, an artificial intelligence (AI) chatbot developed by OpenAI, is increasingly being used in healthcare, including dentistry, for patient education; this study aimed to assess the usability and quality of ChatGPT’s responses to pregnancy-related oral health queries in Saudi Arabia.

Method

This two-part cross-sectional study assessed pregnant Arabic women’s perceptions of ChatGPT for oral health queries and evaluated its responses using an online questionnaire. Responses from ChatGPT-4o mini were rated by 5 dental experts with regard to accuracy, clarity, relevance, and acceptance using a 5-point Likert scale.

Results

Among the 300 participants, 42.0% (126) knew about ChatGPT, 33.7% (101) had previously used it, 14.3% (43) had used it to obtain medical information, 8.7% (26) had used it for dental information, and 8.3% (25) had used it for dental information during pregnancy. Attitudes regarding ChatGPT were rated from 1 to 4. Except for 1 item, the means were all above the midpoint. Attitude ratings ranged from a mean of 2.71 (SD 0.76) for ChatGPT competency to a mean of 2.34 (SD 0.92) for its ability to replace human interactions. However, ChatGPT competency (P = .028), security (P = .015), willingness to use ChatGPT for inquiries (P = .021), ability to assist in informed decision-making (P = .01), willingness to make decisions based on recommendations (P = .024), and persuasiveness (P = .049) were significantly different based on educational level. Pregnant women with higher levels of education rated these aspects significantly lower than those with a high school diploma or bachelor’s degree.

Conclusion

ChatGPT provided useful oral health information for pregnant individuals, but its responses required revision and supervision by health professionals. Its usage among pregnant women in Saudi Arabia remained low.