AI testing for pharma and life sciences

An assistant that answers from SmPCs and standard operating procedures is tested the way your quality unit would expect: acceptance criteria approved before execution, and a dated report that records every deviation.


AI testing for pharma and life sciences. An assistant that answers from SmPCs and standard operating procedures is tested the way your quality unit would expect: acceptance criteria approved before execution, and a dated report that records every deviation.

What can go wrong

An invented example of how it can fail: A production operator asks how long a cleaned vessel may be held before its next use. The assistant answers 72 hours and cites the previous version of the cleaning procedure. Since the last cleaning validation, the current version sets 48 hours. The citation looks right because the document exists, and the deviation surfaces in an audit or a batch investigation.

Questions the test would include:

What the AI Act means here

A document assistant working on SmPCs or internal procedures is not in Annex III and is generally not high-risk. If it answers on the treatment of individual patients, its purpose may make it a medical device under Regulation (EU) 2017/745. If its conformity assessment also requires a notified body, it is high-risk (Article 6(1) and Annex I).

In any case, if the assistant talks with people, they must be told they are talking to an AI unless that is obvious (Article 50, from 2 August 2026). And whoever develops it or uses it in their organisation must take measures to support the AI literacy of the staff who work with it (Article 4). Whether a system is high-risk is a legal question for your adviser. I do not classify it; the test gives you evidence either way.

Criteria approved before execution and every deviation recorded: the test follows the logic your quality unit already applies, without being the GxP validation of the computerised system.