AI changes the evidence pattern, not the need for control.
FDA CSAAI-enabled QMS software
If AI supports quality decisions, intended use and risk drive the assurance file.
Automation confidence
Automated tests or AI-assisted review should have fit-for-use rationale.
Change control
Model, rule, workflow, or vendor changes need impact assessment.
Human review
Critical quality decisions should show who reviewed outputs and what evidence they checked.
AI governance must connect to existing regulated systems.
not isolatedSoftware inventory
Connect this evidence to QMS, clinical, software, supplier, data, or lifecycle governance where applicable.
Intended-use statement
Connect this evidence to QMS, clinical, software, supplier, data, or lifecycle governance where applicable.
Risk assessment
Connect this evidence to QMS, clinical, software, supplier, data, or lifecycle governance where applicable.
Assurance plan
Connect this evidence to QMS, clinical, software, supplier, data, or lifecycle governance where applicable.
Testing rationale
Connect this evidence to QMS, clinical, software, supplier, data, or lifecycle governance where applicable.
Objective evidence record
Connect this evidence to QMS, clinical, software, supplier, data, or lifecycle governance where applicable.
Current public sources for FDA CSA.
official firstThese links are the public source anchors for this workspace. Interpretation, checklists, and future assets should point back here before being reused outside iFeed.
FDA CSA guidance
2026-02 · Current final guidance on computer software assurance for production and quality management system software; it supersedes the September 2025 final guidance.
FDA CSA guidance PDF
2026-02 · Guidance history notes February 2026 revisions to align terminology with QMSR amendments to Part 820.