Participants in a webinar hosted by the Atlantic Council’s Cyber Statecraft Initiative on Sept. 8 discussed the challenges of deploying AI in healthcare. Experts noted that conflicting global regulations regarding data use and cross-border transfers hinder pharmaceutical research and clinical trials.
Ranjit Kumble, VP of enterprise data science at Pfizer, highlighted that international companies must navigate disparate legal frameworks in China, the European Union, and the U.S. In Europe, developers must comply with both the GDPR and the EU AI Act, while U.S. organizations face a fragmented landscape of state and sector-specific privacy laws.
This regulatory fragmentation creates scalability issues, as companies often must prepare datasets separately for each project to ensure compliance. Such repetitive work is described as unsustainable, potentially impacting the quality of AI models and a company's ability to compete in the market.
The discussion also addressed the risks associated with agentic AI systems. Daria Bahrami of Dreadnode and Stephen Moon of Snowflake emphasized that as AI agents gain the ability to act with limited human oversight, organizations must implement strict access controls, data segmentation, and continuous monitoring to prevent security breaches.
Experts advise healthcare organizations to proactively manage these risks by creating inventories of AI use cases, documenting pre-deployment reviews, and establishing flexible governance structures. These measures are intended to help companies adapt to jurisdiction-specific requirements without rebuilding their compliance programs for every new project.