Scholarlink

Purpose <p>This scoping review aims to critically assess the applications, performance, and limitations of artificial intelligence (AI) chatbots in oral pathology.</p> Methods <p>An electronic search was conducted in PubMed/MEDLINE, Embase, Scopus, and Science Direct from inception to March 2025. Studies were eligible if they evaluated chatbot use in oral pathology, including diagnostic support for oral lesions and cancers, patient education, and professional training.</p> Results <p>From 761 records identified, 13 studies met the inclusion criteria. AI chatbots demonstrated potential to assist in diagnostics and patient education on oral diseases. Diagnostic accuracy was influenced by lesion type, patient characteristics, and whether multimodal inputs (e.g., images plus text) were provided, with higher accuracy observed when such data were included. Despite these strengths, chatbots frequently exhibited limitations such as inconsistent empathy, citation inaccuracies, and variable readability—issues that were recurrent across studies and highlight the need for professional oversight.</p> Conclusion <p>AI chatbots show promise in supporting diagnostics and patient education in oral pathology. However, key limitations regarding accuracy, empathy, and information reliability necessitate robust validation and ethical safeguards before clinical adoption.</p> Prospero registration <p>CRD420251020464.</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

Artificial Intelligence chatbots in oral pathology: a scoping review of applications, accuracy, and limitations

  • 期刊论文

  • 发表日期：

    2025年11月16日

查看全文
  • Everton Freitas de Morais,
  • Maurília Raquel de Souto Medeiros,
  • Deborah Gondim Lambert Moreira,
  • Lilianny Querino Rocha de Oliveira,
  • Roseana de Almeida Freitas

摘要

Purpose

This scoping review aims to critically assess the applications, performance, and limitations of artificial intelligence (AI) chatbots in oral pathology.

Methods

An electronic search was conducted in PubMed/MEDLINE, Embase, Scopus, and Science Direct from inception to March 2025. Studies were eligible if they evaluated chatbot use in oral pathology, including diagnostic support for oral lesions and cancers, patient education, and professional training.

Results

From 761 records identified, 13 studies met the inclusion criteria. AI chatbots demonstrated potential to assist in diagnostics and patient education on oral diseases. Diagnostic accuracy was influenced by lesion type, patient characteristics, and whether multimodal inputs (e.g., images plus text) were provided, with higher accuracy observed when such data were included. Despite these strengths, chatbots frequently exhibited limitations such as inconsistent empathy, citation inaccuracies, and variable readability—issues that were recurrent across studies and highlight the need for professional oversight.

Conclusion

AI chatbots show promise in supporting diagnostics and patient education in oral pathology. However, key limitations regarding accuracy, empathy, and information reliability necessitate robust validation and ethical safeguards before clinical adoption.

Prospero registration

CRD420251020464.