What it is

FDA issued this draft guidance on 7 January 2025 (docket FDA-2024-D-4689), jointly from CDER, CBER, CDRH, CVM, and other offices. It covers AI used to produce information or data that supports regulatory decisions on safety, effectiveness, or quality. As of October 2026 it remains a draft.

The 7-step credibility framework

  1. Define the question of interest.
  2. Define the context of use (COU).
  3. Assess AI model risk.
  4. Develop a credibility assessment plan.
  5. Execute the plan.
  6. Document the results and any deviations.
  7. Determine whether the model is adequate for the COU.

Model risk combines model influence (how much the AI output contributes relative to other evidence) and decision consequence (how serious an adverse outcome from a wrong decision would be).

Out of scope: drug discovery, and operational uses that do not affect patient safety, drug quality, or study reliability, such as drafting submissions.

  • FDA held a public workshop on AI guiding principles on 7 October 2025.
  • In January 2026, FDA (CDER and CBER) and EMA issued ten Guiding Principles of Good AI Practice in Drug Development.
  • CDER reports over 500 submissions with AI components from 2016 to 2023.

Reuse terms

FDA website content is in the public domain unless otherwise noted and may be republished freely; credit is appreciated, and FDA suggests noting the date copied. Links to other sites are not endorsements. Third-party content keeps its rights.

Using the draft in AI and RAG

  1. Tag it as Draft with its date and docket.
  2. Use it for planning, not as current requirements.
  3. Record COU and risk level for each AI model that feeds regulatory evidence.
  4. Link to the EMA-FDA principles as a separate document.
  5. Note the scope exclusions: a RAG tool drafting a submission falls outside it, but other rules still apply.

Common mistakes

  • Treating the draft as final.
  • Applying it to drug discovery uses that are out of scope.
  • Skipping the COU definition.
  • Forgetting that operational AI still needs validation under GMP and Part 11 where relevant.

Checklist

  • Draft status and date recorded
  • COU defined per model
  • Model influence and decision consequence assessed
  • Credibility plan documented
  • EMA-FDA principles linked

How AskDona helps

AskDona checks your AI model documentation against the seven steps and shows where each step is evidenced, with citations to your files and the guidance.

Run a sample AI credibility plan check

AskDona shows the evidence; your regulatory and quality teams decide.

Editor’s note. This article is general information, not legal advice. The “conditions to check” rating comes from our source list and is a first-pass judgment, not a legal opinion.

FAQ

Is FDA’s AI guidance for drugs final? No. It was issued as a draft on 7 January 2025 and remained a draft as of October 2026.

What is a context of use? The specific role and scope of the AI model in answering the question of interest.

Does it cover AI used to write submissions? No. Operational uses that do not affect safety, quality, or study reliability are out of scope.

What did FDA and EMA publish in 2026? Ten Guiding Principles of Good AI Practice in Drug Development, in January 2026.

References

All pages opened and checked by October 26, 2026.

SourceWhat it supports
FDA: Draft guidance pageDraft status, docket, issuing centers
FDA: Draft guidance text (PDF)Seven steps, model risk, scope
FDA: CDER AI in drug developmentStatus, workshop, guiding principles
FDA: Website policiesPublic domain reuse
FDA: Website disclaimerLinks not endorsements