MCKEESPORT, Pa. — Adam Baughman, a student in the biomedical engineering technology program at Penn State, joined instructor Joie Marhefka in presenting original research on artificial intelligence (AI) at the 2026 AAMI eXchange, held June 19-22 in Denver, Colorado. Another biomedical engineering technology student, Braxton Marshall, also contributed to the project but did not attend the conference.
Hosted by the Association for the Advancement of Medical Instrumentation, AAMI eXchange is a leading conference for healthcare technology management professionals. The event brings together clinical engineers, biomedical equipment technicians, healthcare technology leaders and industry partners from across the United States and around the world.
The independent research project, titled “AI in HTM: Troubleshooting and Beyond,” grew out of Marhefka’s work exploring artificial intelligence through Penn State’s Faculty Academy, combined with ideas developed by Baughman.
The project consisted of two components. The first surveyed healthcare technology management professionals to better understand how artificial intelligence is currently being used in the field, its potential future applications, challenges associated with adoption and the AI-related skills employers expect from prospective employees.
The second evaluated the effectiveness of several generative AI platforms in troubleshooting medical device issues. Researchers compared responses from Perplexity, ChatGPT, Grok, Gemini and Copilot against manufacturer documentation.
The researchers found that awareness of artificial intelligence is widespread among healthcare technology professionals and that AI is already being used for administrative tasks and some troubleshooting and preventive maintenance procedures. Respondents also identified predictive maintenance as a promising future application.
At the same time, the research highlighted several challenges. Many professionals reported not yet using AI in their work, while common concerns included inaccurate information, over-reliance on AI and data security. The study also found that employers increasingly value applicants who possess a basic understanding of artificial intelligence and effective prompt writing.
When evaluating the AI platforms, the researchers found that no single tool consistently outperformed the others. While each platform could provide basic information, many responses required additional questioning to verify accuracy, and some generated incorrect information or inactive reference links. Based on the findings, the researchers concluded that generative AI can serve as a useful starting point but should not replace authoritative sources when troubleshooting medical devices.
Marhefka said presenting at AAMI eXchange gave Baughman the opportunity to share his work with an international audience of healthcare technology professionals.
Conference attendees responded positively to the presentation, Marhefka said, noting that student presenters are uncommon at the event. Discussions following the session centered on the advantages of proprietary artificial intelligence systems compared with publicly available platforms, as well as the AI knowledge and skills employers hope to see in future healthcare technology professionals.