As healthcare is being transformed with AI, different global regulators are implementing diverse strategies to ensure both new developments and patient safety. EU members are following a full risk assessment process, the United States is revising its medical device rules supporting AI and Japan is choosing a sandbox built around flexible guidelines to encourage progress while respecting privacy.
Table of Contents
EU’s AI Act: Setting the Gold Standard
The AI Act established by the EU now sorts AI applications by assigning them to three categories: those that are prohibited, those that are high-risk and those that are low-risk. Because medical devices with AI are classified as “high risk,” they require careful adherence to standards such as risk management, managing data, and providing human supervision . Extra requirements were added and they are required in addition to the current EU Medical Device Regulation and In Vitro Diagnostic Regulation.
FDA’s Lifecycle Approach: Balancing Innovation and Oversight
The U.S. Food and Drug Administration (FDA) is now using a flexible and repetitive method for regulating AI in healthcare. In December 2024, the FDA issued final guidance on using PCCPs for medical devices that use AI. With PCCPs, manufacturers are able to improve their AI systems on the go by applying eligible updates and saving time with new marketing requests.
Japan’s Regulatory Sandbox: A Safe Space for Medical AI Innovation
AI technology in healthcare is being supported in Japan thanks to a policy called the Regulatory Sandbox Framework which was established in 2018. As part of the initiative, companies can explore and use new solutions with exemptions from regulations that are only temporary. The purpose of using a sandbox is to accelerate the development and use of new inventions in healthcare.
Global Lessons: One Technology, Three Visions
These three territories use various approaches to include AI in healthcare. Prioritising strict rules and responsibility for safety in the EU, might decrease the rate of innovation. The US uses a flexible method, supporting ongoing changes as it keeps an eye on the system. By using its sandbox model, Japan is working on ways to solve data privacy issues while supporting new technologies.
While AI revolutionises medicine, these different procedures for regulating AI reveal useful advice for promoting both progress and safety.
