Related Experiment Video
Updated: Jan 7, 2026
05:33
Introduction of an Integrated Pathology Image Management, Artificial Intelligence, and Reporting System
Published on: July 11, 2025
741
The AI-powered pathologist: A global survey mapping initial trends in AI adoption and outlook
Meredith K Herman1, Sania Qazi2, Elisa Farrell2
1University of Michigan Department of Pathology, Ann Arbor, MI, United States of America.
Journal of Pathology Informatics
|December 29, 2025
Summary
Pathology professionals are familiar with artificial intelligence (AI) but cautious in its adoption, primarily using tools like ChatGPT for non-diagnostic tasks. Limited clinical use and institutional support highlight the need for guidelines and education.
Area of Science:
- Medical Informatics
- Digital Pathology
- Artificial Intelligence in Medicine
Background:
- Artificial intelligence (AI) tools, including large language models (LLMs) and vision-language models (VLMs), are rapidly emerging and impacting various professional fields.
- Pathology, a field reliant on complex data interpretation, stands to be significantly influenced by AI integration in training and clinical practice.
Purpose of the Study:
- To assess the adoption, usage patterns, perceptions, and challenges of AI-driven tools among pathology trainees and practicing pathologists.
- To explore the current integration of AI in pathology training and clinical workflows and identify future directions.
Main Methods:
- A cross-sectional, anonymous online survey was distributed to pathology residents, fellows, and attending pathologists globally.
- The survey gathered data on AI familiarity, usage frequency, perceived benefits and risks, and institutional policies using multiple-choice, Likert-scale, and open-ended questions.
- Data were analyzed using descriptive and inferential statistics, with qualitative responses undergoing thematic categorization.
Main Results:
- Out of 268 respondents from 23 countries, 73% reported AI familiarity, but actual usage was limited (31% rare use, 29% no use).
- ChatGPT was the most utilized tool (84%), primarily for document drafting, research, and administrative tasks, with minimal diagnostic application.
- Key concerns included accuracy (81%), over-reliance (65%), and data security (63%), with only 10% reporting clear institutional AI guidelines.
Conclusions:
- AI is increasingly adopted in non-diagnostic pathology tasks, but its integration into clinical practice and training remains cautious.
- Significant gaps exist in AI's clinical application, user trust, and institutional support within pathology.
- Developing clear guidelines, targeted education, and robust validation are crucial for the safe and effective integration of AI in pathology.
