Generative artificial intelligence (AI) describes AI that creates new content, including text, videos, images, and audio. The reaction to AI in the medical community has ranged from horror and predictions of doom and gloom to true excitement.
The current general consensus seems to be cautious optimism about the potential benefits, with the caveat that there need to be appropriate guardrails.
Thus, it is timely that this week, Pediatrics is early releasing a State of the Art Review article about the potential uses and pitfalls of AI—specifically large language models (LLM), such as ChatGPT—in the realm of pediatric education. This article by Dr. Srinivasan Suresh from the University of Pittsburgh and Dr. Sanghamitra Misra from Baylor College of Medicine is entitled, “Large Language Models in Pediatric Education: Current Uses and Future Potential” (10.1542/peds.2023-064683).
The authors provide examples of the potential of LLM in supporting teaching and improving patient care:
- Creation of tools to assess skills
- Development of tools to provide training in specific topics, e.g., how to talk to adolescents about sensitive topics
- Improved written feedback to learners
There is also great potential in improving patient care. For instance, LLM can provide written patient education in the patient’s native language, and this can easily be adjusted to the child’s age and the literacy level of the family.
The authors also thoughtfully consider the limitations, challenges, and ethical considerations of using LLM in pediatric education. These include but are certainly not limited to:
- The potential inaccuracy of the information generated
- Overreliance on LLM at the expense of developing critical thinking skills
- The increased potential for plagiarism
I have only skimmed the surface of all that is in this review. AI is not going away, and we all need to consider how we can maximize the benefits of AI while optimizing the education of our trainees and the care of our patients. This State of the Art article is important reading for all of us.
