Adoption of AI Outpaces Training in Field Epidemiology Programs, Survey Finds

A new survey has revealed that two-thirds of epidemiologists in specialized training programs are using artificial intelligence (AI) in their work, despite only one-fifth having received formal or informal AI training. The study, published in Eurosurveillance, highlights a significant gap between the adoption and training of AI in field epidemiology programs.

The survey was conducted by researchers from the Public Health Agency of Canada among 105 participants in Field Epidemiology Training Programs (FETPs), also known as field epidemiologists. These specialized programs train individuals to develop technical capabilities in applied epidemiology, improve their professional judgment, and enhance their ability to communicate information effectively.

FETPs are designed to equip professionals with the skills necessary for real-world applications of epidemiological principles. However, the increasing reliance on AI raises concerns about the development of critical-thinking skills and professional discernment in a field that places a premium on judgment and interpretation, as well as the ability to communicate uncertainty.

The researchers note that while AI has the potential to improve efficiency for field epidemiologists, it also poses challenges related to the development of these essential skills. The study’s findings underscore the need for comprehensive training programs that address both technical and professional aspects of AI adoption in field epidemiology.

Among the 105 respondents, 66% reported using AI in their work. This figure was highest among fellows in the European program (89%), followed by those in the US program (54%) and Canadian program (54%). Notably, there was no statistically significant difference between first- and second-year fellows regarding AI adoption.

The survey also explored how frequently respondents used AI tools. Among users, 42% reported using AI weekly, while 30% said they use it daily, and 26% occasionally. Most respondents felt either somewhat comfortable (49%) or very comfortable (35%) when using AI, with only a small percentage feeling uncomfortable.

The most commonly used platform was ChatGPT (87%), followed by institution-specific tools (25%). The primary applications of AI were troubleshooting coding errors (91%), writing code (75%), and improving work efficiencies (42%). Qualitative responses suggested that AI generally improved coding efficiency, reducing the time spent on error resolution and facilitating learning new techniques.

Respondents also reported using AI to improve writing quality and streamline routine administrative tasks. Some noted that AI helped them better understand complex topics, conduct background research, and summarize information effectively. However, a significant proportion of users (41%) reported barriers to adopting AI technology, citing technical limitations, limited access to tools, uncertainty about institutional rules governing AI use, and concerns regarding accuracy and reproducibility.

Ethical considerations were also a concern for one-quarter of AI users, including data privacy, potential bias, environmental impacts, and the risk of overreliance on AI. Respondents emphasized the need for practical instruction on using AI for coding, data analysis, scientific writing, data visualization, and outbreak detection, as well as training on ethical considerations and effective prompting strategies.

The survey highlights substantial gaps in institutional guidance and training regarding AI adoption. Only one-fifth of fellows reported receiving any formal or informal AI training. Respondents expressed interest in learning about the limitations of AI, such as increased awareness of common mistakes and potential pitfalls.

In light of these findings, researchers recommend that FETP curricula promote AI literacy by incorporating education on the technology’s limitations, biases, and ethical implications. Possible recommendations for training might include hands-on workshops, real-world case studies, and access to AI tools and platforms.