Task Force Meeting 12

September 24, 2025

Quick recap

The meeting began with an exploration of AI-related developments and challenges in research and teaching. The team discussed potential keynote speakers for an upcoming AI Research Day and considered various topics including digital twins, grant writing, and AI policy implementation in public health. They concluded by addressing the limited adoption of AI in public health compared to medicine and discussed plans for future events and collaborations.

Next steps

• Michelle to look into available resources for AI detectors and provide recommendations from IT on how to use them.

• Beth to share AI and grant writing resources with Junie for collation.

• Junie to send AI and grant writing resources to the entire PHAISE Task Force.

• Junie to create a wiki page for best practices for AI and grant writing.

• PHAISE Task Force to continue identifying potential keynote speakers for AI Research Day in November/December or January.

• PHAISE Task Force to reach out to the RAISE initiative from the Department of Medicine and invite them to a future meeting.

• Junie to ensure Carl is added to the data discussion and commercialization of data group.

Summary

AI Research and Applications Discussion

The group discussed AI-related developments and challenges, with Carl expressing interest in exploring AI's potential for survey research and data analysis, while Kadija highlighted the need for updated ethics frameworks for large language models. Gregory mentioned ongoing qualitative research in Uganda and the challenges with using AI tools for analysis. The team also discussed the Hopkins AI Hub, now rebranded as HopGPT, and its accessibility issues. Michelle, representing the staff assembly, shared updates on using AI to summarize course evaluation results, noting some concerns about data representation and interpretation. The conversation ended with plans to organize the next AI Research Day, tentatively scheduled for November, with a focus on increasing participation and showcasing AI-related work.

AI Research Day Speaker Selection

The team discussed potential keynote speakers for AI Research Day, initially considering Alain Labrique from WHO but decided to explore other options. Kadija suggested Sanjay Gupta, a former CNN correspondent and neurosurgeon with academic experience, though Gregory expressed concerns about his current clinical practices. The group emphasized the need for a speaker who could provide a global health perspective on AI applications, differentiating it from healthcare AI discussions. They agreed to review last year's speaker list and consider Sanjay Gupta's availability, with the possibility of moving the event from November to February if needed.

Digital Twins and AI Research

The group discussed the concept of digital twins, exploring their applications in medicine and public health, though they decided against using a population-scale digital twin speaker for a seminar due to lack of fit with their criteria. Gregory announced plans to organize an AI Research Day with a format similar to previous events, including abstract submissions and awards. The group also addressed a new NIH policy limiting the number of grants Principal Investigators can submit annually to six, with Gregory expressing support for this restriction as it promotes quality over quantity in grant submissions.

AI in Grant Writing Concerns

The group discussed concerns about AI usage in grant preparation, particularly regarding NIH's stance on original content and AI detection tools. Beth shared her work on a fellowship grant writing workshop and mentioned guidelines for appropriate AI use in academic contexts. Gregory inquired about available AI detection resources, and Michelle agreed to look into it. The team considered surveying recent NIH grant applicants about their AI usage, though Gregory expressed hesitation about the response rate. They also discussed ASPPH's ongoing work on synthesizing AI-related policies across public health schools, noting a lack of specific guidance for research applications.

AI Policy for Research and Teaching

The group discussed the implementation of a new AI policy for research and teaching, noting that while there is a university-wide committee on AI and teaching, no specific guidance exists yet for grant writing and research. Beth shared resources on using AI for research, including a module on grant writing, and offered to collaborate on creating an internal curated resource. The group agreed to compile additional resources and share them on the research wiki, with a focus on using AI as a critical reflective tool for grant writing while ensuring confidentiality. They also discussed the importance of proper prompt engineering to get better AI responses and the need to use secure, protected environments for AI assistance.

AI Adoption in Public Health

The group discussed the limited adoption of AI in public health compared to medicine, attributing it to factors such as lack of creativity, network connections, and the nature of public health questions being less focused. Khadija proposed having Arjun Menry, a data scientist and Deputy Editor of NEJM AI, speak on precision public health at an upcoming event. Gregory suggested continuing to develop the event for November or early December, with a backup plan for January or February if not enough progress is made. He also shared updates on two AI-related activities in the Department of Medicine and the Collaboration Center on AI and Aging, expressing interest in potentially collaborating with these groups.

Recording

PHAISE Task Force Meeting 12 Recording.mp4