AI fundamentally shifts software testing in medical devices from deterministic verification to data-driven, probabilistic validation.
It requires integration into a continuous PDCA cycle, combining black-box testing, statistical evaluation, and risk management.
In this session we will discuss how to validate AI of unknown provenance, where safety can be ensured through strict use limitations and continuous post-market monitoring rather than complete pre-market verification.
Presented by: Regina Geierhofer, Chair of CLC TC62, Regulatory Affairs Professional
This recording was broadcasted during an Aligned Sharpen your Skills event.
Start a free trial: https://www.aligned.ch/free-trial
Visit our website: https://www.aligned.ch
Download
0 formats
No download links available.
Testing the untestable: Computer Software Validation of AI functionality | NatokHD