The US Food and Drug Administration (FDA) has finalized a long-anticipated guidance on using predetermined change control plans (PCCPs) in submissions for medical devices with artificial intelligence (AI) features. The finalized guidance clarifies the agency’s thoughts on AI versus machine learning (ML) products and adds several sections, including a section on version control of PCCP-enabled devices.
On 3 December, FDA finalized the guidance, which details its expectations for allowing PCCPs for AI-enabled devices. Regulators reiterated that ML is a subset of AI. However, while the draft version focused on providing a forward-thinking approach to ML development, the final version takes a broader approach and says it is intended for AI-enabled device software functions (AI-DSF) more generally…
