Pulmonary Embolism Education: Role of Generative Artificial Intelligence Models

Area of Science:

• Health Informatics
• Artificial Intelligence in Healthcare
• Patient Education

Background:

• Generative artificial intelligence (AI) is increasingly used for information dissemination.
• Improving patient health literacy is crucial, particularly for complex conditions like acute pulmonary embolism.
• Existing literature explores AI's potential in enhancing patient understanding of diseases and management.

Purpose of the Study:

• To evaluate the readability of generative AI responses for patient health information.
• To determine if current AI models meet established US readability recommendations.

Main Methods:

• Assessed the readability of responses generated by publicly available AI models (ChatGPT, Google Gemini, Microsoft CoPilot).
• Compared AI-generated content against United States readability standards.

Main Results:

• ChatGPT, Google Gemini, and Microsoft CoPilot responses did not meet current US readability recommendations.
• The generated content may pose challenges for patient comprehension.

Conclusions:

• Publicly available generative AI models currently produce health information that is not easily readable for patients.
• Longitudinal studies are necessary to track the evolving readability of AI models as they adapt through deep learning.