Session 3: Artificial Intelligence - NPHIs Role in Utilizing the Opportunities and Addressing the Risks
On April 18, 2024, IANPHI held a session of its Europe Regional Network meeting on “Artificial Intelligence - NPHIs Role in Utilizing the Opportunities and Addressing the Risks”.
Under the moderation of Dr. Trygve Ottersen, Executive Director of the Norwegian Institute of Public Health (FIH) and Chair of the IANPHI Europe Regional Network, who highlighted the significance of discussing artificial intelligence (AI)'s impact on society and institutions, the session delved into European Union (EU)’s regulatory frameworks on AI with the presentation of Dr. Yiannos S. Tolias, Legal Lead AI and AI Liability in Healthcare of the European Health Data Space (EHDS) Team at the European Commission (EC). Dr. Ottersen also presented insights from a recent survey on AI use in National Public Health Institutes (NPHIs), stressing the diversity of AI applications alongside shared challenges, and the importance of collaboration in navigating AI implementation and governance, which was followed by a table discussion.
EU’s Regulatory Frameworks on AI and its Implications for Public Health
By Dr. Yiannos S. Tolias, Legal Lead AI and AI Liability in Healthcare, European Health Data Space (EHDS) Team, European Commission (EC)
Presentation of the Survey and of the Outline for a Framework For how NPHIIs can Approach Artificial Intelligence
By Dr. Trygve Ottersen, Executive Director, Norwegian Institute of Public Health (FIH), Norway, and Chair, IANPHI Europe Regional Network
Dr. Ottersen provided an overview of the results from the recent survey conducted by IANPHI Europe Regional Network on the use of AI within the National Public Health Institutes (NPHIs) and the subsequent analysis of the responses. He presented a draft outline of a framework for how NPHIs can address AI, and this outline was discussed in groups during the session.
In the introduction, Dr. Ottersen also highlighted some of the activities undertaken by IANPHI so far, including a webinar hosted by IANPHI Europe last October and a special session during the IANPHI Annual Meeting in Kigali on AI. Recognizing the importance of sharing insights and learning from other institutes' experiences, a survey was disseminated to members from the IANPHI Europe Regional Network, yielding responses from 16 institutes. The survey encompassed questions pertaining to current AI applications, experiences, planned measures, and areas where support from IANPHI might be beneficial. Dr. Ottersen emphasized the variability in responses, especially on the total amount and on the specific types of AI applications. Despite potential differences in understanding AI terminology, the survey offered valuable insights into the scope and diversity of AI-related activities.
From the survey, several themes emerged as shared at least among some NPHIs, including text summarization, document search, literature review, surveillance – especially of infectious diseases –, forecasting, image recognition, disease diagnosis and prediction, and text mining. Notably, institutes also showcased unique applications, ranging from predicting nitrate concentration in groundwater to analyzing drug at border crossings. Challenges identified by the institutes included data protection, privacy, competence, bias, inaccuracy, transparency, governance, infrastructure limitations, public understanding, external collaboration, workforce concerns, and ecological impact. Strategies to address these challenges included formulating guidelines, mapping activities, training, fostering external collaborations, recruitment, strategy development, research, and enhancing internal and external roles.
In conclusion, Dr. Ottersen highlighted the diversity of AI applications across institutes, alongside shared challenges and potential avenues for collaboration and support. He underscored the importance of collective action and knowledge exchange in navigating the complex landscape of AI implementation and governance.
Table Discussion
Dr. Ottersen invited participants to engage in table discussions to explore the potential usefulness and structure of a framework. Following the table discussions, participants reconvened to share their insights. Several key points were raised, including the need to characterize specific issues and questions relevant to public health institutes, such as ethical considerations, legal aspects, and barriers to technology access. Participants suggested incorporating sections within the framework to address these concerns, emphasizing the importance of shared experiences and best practices in public health applications. Ethical dimensions, regulatory frameworks, and the development of use case identification were identified as essential components to consider. The importance of collaboration, information sharing, and standardization efforts at both national and EU levels were emphasized.
Moving forward, participants agreed on the value of further discussions and possible webinars to delve deeper into specific topics, including a webinar on the successful applications of AI in NPHIs. Despite challenges in defining AI, participants expressed a commitment to progress and collaboration in navigating this rapidly evolving field.