Artificial intelligence has been edging into infectious diseases for several years, assisting with documentation, surfacing patterns in data and nudging clinicians toward safer prescribing.
An IDWeek premeeting workshop, “Bug Meets Bot: Practical AI for Infectious Disease,” Oct. 21, 8 a.m. to noon ET, will help ID professionals better understand AI’s potential and how to use it thoughtfully.
“AI has graduated from being a tool to consider in the future to a present-tense feature of our jobs,” said Chidi Akusobi, MD, PhD, an infectious diseases and clinical informatics fellow at Stanford Health Care and one of the organizers of the workshop. “For ID, we feel the question has shifted from ‘Will this matter?’ to ‘How do we in ID shape it, wield it, and make sure it works for our patients and our field?’”
A Practical Introduction to Current Tools
The organizers built the premeeting workshop with one goal in mind: Give ID clinicians a practical, hands‑on introduction to AI tools they can use right now. It was designed to cover the gamut of experiences, too, said workshop speaker Nicholas Paul Marshall, MD, FAAP, a clinical fellow in clinical informatics at Stanford University who also trained in pediatric infectious diseases.
“For those newer to AI, we start from the beginning. The story of how we got here, how these tools actually work and how to get excellent results from them,” Dr. Marshall said. “For those already using AI regularly, the field is moving fast enough that even experienced users will find value.”
The workshop also will include plenty of time for discussion, he said. “And we expect the conversation between beginners and experienced users to be one of the most valuable parts of the day.”
Three recent advances make now the right time to explore AI more deeply, Dr. Akusobi said. AI tools have become “widely adopted in clinical care,” he said, citing a survey from the American Medical Association that found about 81% of physicians are using an AI tool. “The ubiquity of these tools means that we can no longer ignore their impact on how ID physicians practice.”
The Growth of Agentic AI
The advances in AI models — particularly in clinical reasoning — and the advent of agentic AI systems add additional urgency. “AI can now execute multistep tasks in the real world, and it’s only a matter of time before agentic AI meets clinical practice,” Dr. Akusobi said. “This is a categorical change. AI is poised to move from a tool physicians use to augment decisions or perform clerical work to a tool that does the work of physicians,” he said. “Given this likely future, it’s important that physicians are empowered with this knowledge.”
Dr. Marshall sees ID clinicians already adopting AI tools to “quickly find answers during their consults or to help with consultations. Those tools are valuable on their own, but what really excites me is what comes next.” He sees a role for agentic AI to automate recurring and time-consuming tasks without coding.
This could be “something as straightforward as an agent that takes raw antimicrobial stewardship data and automatically transforms it into a polished report with figures, tables and trend analyses,” Dr. Marshall said. “Or something more sophisticated, like an agent integrated with the electronic health record to assist with complex dosing decisions. The research applications are enormous.”
Addressing Concerns
Still, there are concerns around use of AI tools related to privacy and safety, particularly with some of the widely available tools.
“Any time we input data into a large language model (i.e., like Claude, Gemini, ChatGPT, etc.), we are allowing a third party visibility to the questions we ask and the context we provide for each prompt,” said workshop speaker Aydin Zahedivash, MD, an adjunct clinical assistant professor in clinical informatics at Stanford. “The only HIPAA-compliant way to use patient information with any form of online AI is if your health institution has a relationship with the AI company in the form of a business associate agreement, which outlines how patient data should be handled when used with their tool.”
Both Drs. Akusobi and Marshall use AI in their own work. Dr. Akusobi uses large language models to look up information or “obtain a second opinion on my diagnostic reasoning for difficult cases.” They may summarize cases when transferring service to a new provider or compile a patient’s microbiology history.
Dr. Marshall has moved “from user to builder,” including tools “that I once thought would require a dedicated engineering team.” He’s recently completed a vaccine app — built in an afternoon without any formal software development background.
“We’re at a moment where a clinician with a good idea and a willingness to experiment can go from concept to working prototype faster than ever before.”
Learn more about all the premeeting workshops at IDWeek and how to register.