P2.068: Assessing Attitudes Toward Artificial Intelligence in Medicine Among Coastal Virginia Pediatric Residents

Background: A search for "AI" on Google often leads to OpenAI’s website, the creator of ChatGPT, a tool widely recognized by adults in the United States. However, AI’s influence extends far beyond ChatGPT, seamlessly integrating into daily life through devices such as Fitbit fitness trackers, Nest and Ring security cameras, and email spam filters. Other notable examples include virtual assistants like Apple’s Siri and Amazon’s Alexa. As AI becomes more prevalent, attitudes, perspectives, and opinions regarding the technology continue to evolve. This study aims to assess the perspectives of pediatric residents in Coastal Virginia on the use of AI in medicine, a particularly relevant issue as AI starts to be integrated into medical systems.

Methods: The General Attitudes towards Artificial Intelligence Scale (GAAIS), developed in 2020, was adapted (with permission from the original creators) to assess the perspectives of physician trainees, particularly pediatric residents. The survey (EVMS IRB #24-08-XX-0202-EVMS) was distributed to pediatric residents between 8/24 and 3/25. The survey (Fig. 1) used a 5-point agreement Likert scale. Responses were analyzed to identify differences based on training levels. Survey questions were divided into positive and negative subscales. A two-sided chi-squared test was used for analysis. All tests performed using RStudio (2-sided, p< 0.05 = statistically significant).

Results: Of the 84 pediatric residents at EVMS/ODU, 44 responded to the survey, resulting in a 52.4% response rate (Fig. 2a). The level of all agreement with positive statements about AI use in healthcare was significantly higher across all training levels as compared to negative statements (p < 0.001, 95% CI (17.6, 32.3)) (Fig. 2b). Notably, PGY3s were more likely to disagree with the positive applications of AI and express agreement with negative views compared to PGY1s (p=0.005, 95% CI (5.5, 32.1)) and PGY2s (p=0.019, 95% CI (2.7, 29.9)) (Fig. 2c and 2d). Furthermore, PGY2s exhibited a sharp increase in positive attitudes toward AI compared to PGY3s (p=0.018, 95% CI (3.1, 27.9)) (Fig. 2c and 2d).

Conclusion: The survey suggests that, regardless of training level, pediatric residents in Coastal Virginia generally hold positive attitudes toward AI in medicine. However, PGY3s demonstrate more skepticism toward AI’s use in medicine compared to PGY1s. It would be insightful to explore whether PGY1s' attitudes toward AI evolve as they advance in their training, potentially aligning more with the perspectives of current PGY3s. Additionally, investigating whether this trend holds true across other subspecialties could reveal whether attitudes toward AI become more negative as one progresses through training. As AI continues to expand, it is essential to consider and evaluate residents’ attitudes toward its use in healthcare before further integrating AI into medical systems.