The University of North Dakota School of Medicine and Health Sciences is actively considering the establishment of a dedicated artificial intelligence hub aimed at enhancing healthcare in North Dakota. This initiative is intended to support both healthcare professionals and aspiring medical practitioners.
Richard Van Eck, who holds the Monson Endowed Chair for Medical Education and serves as the associate dean for teaching and learning at the institution, shared that discussions about incorporating AI into the medical curriculum began approximately two years ago through informal meetings.
“We realized the necessity of discussing AI, as it was a subject we had limited understanding of. We needed to educate ourselves on it,” he stated. This initial discussion group expanded from just four members to around 40, which led the dean to encourage them to consider future plans for an AI hub.
Van Eck outlined three primary objectives for the proposed AI hub:
1. Collaborating with rural healthcare facilities to develop AI solutions to specific challenges. 2. Establishing partnerships for research initiatives. 3. Providing training to students to enhance their competencies in AI.
The medical school has formed a formal AI committee alongside several advisory groups, each targeting specific focus areas, including policy guidance and learning methodologies. Staff members are currently delving into various dimensions of AI, such as ethical considerations and student apprehensions regarding AI technology, while emphasizing the importance of keeping human input central to AI discussions.
Furthermore, the school is planning literacy training for its students, faculty, staff, and administrators.
“We’ve outlined a strategic plan for the upcoming year,” Van Eck explained. “Our aim is that by the end of this year, we can confidently move forward to invest in building the AI hub, as we will have established a solid foundation. This undertaking will require significant resources.”
Additionally, there is keen interest from the broader Grand Forks community regarding AI’s potential. To address this, Van Eck conducted a community class on June 11, introducing the fundamental concepts of AI as they pertain to healthcare. He clarified differences between generative AI and deep learning while discussing both the advantages and challenges associated with AI technology.
Understanding of AI within the community is quite diverse, he noted, emphasizing the importance of education in dispelling misconceptions about AI.
“The complexity of AI makes it crucial to convey accurate information. We don’t want people to fear or avoid it, as this could impede their growth. The more informed the community is, the better equipped they will be to make sound decisions,” he remarked.
He pointed to a specific AI model developed for pancreatic cancer detection, which demonstrated an impressive 90% accuracy rate in identifying lesions, doing so a year earlier than traditional radiologists, clearly illustrating the impact of timely detection in aggressive cancer forms. Another AI model introduced in 2024 achieved a classification accuracy of between 92% and 97% for distinguishing between cancerous and benign tumors.
In addition, Mayo Clinic has 108 AI models actively deployed in clinical settings, with another 375 currently under development. The Mayo Clinic has also collaborated with Microsoft to create an AI health tool derived from historical patient records.
“Media coverage often focuses on sensational headlines about AI, which can heighten public apprehension,” Van Eck commented. “While these concerns are valid, it’s crucial to highlight the positive potential of AI, especially in solving significant problems and improving lives in ways that might not be possible through other means.”
However, there are still challenges to navigate. For instance, a tool called OpenEvidence, designed to be informed by verified sources like the Mayo Clinic and leading medical journals, mistakenly advised a North Dakota physician to administer a potentially lethal drug dosage to a dialysis patient. This incident underscores the necessity of human oversight in AI applications.
In summary, the University of North Dakota is pioneering efforts to integrate AI into healthcare solutions, balancing technological advancement with ethical and human considerations.
