Pharma AI glossary
AI literacy
Defined in EU AI Act Article 3(56) as the "skills, knowledge and understanding" that allow providers, deployers, and affected persons "to make an informed deployment of AI systems, as well as to gain awareness about the opportunities and risks of AI and possible harm it can cause.
Definition
Defined in EU AI Act Article 3(56) as the "skills, knowledge and understanding" that allow providers, deployers, and affected persons "to make an informed deployment of AI systems, as well as to gain awareness about the opportunities and risks of AI and possible harm it can cause." In the EU it is a legal obligation for organizations (Article 4, since Feb 2025). In the US there is no statutory equivalent, but FDA's entire framework (credibility assessments, CSA, Part 11 review) presumes personnel competent to perform it, so literacy is effectively a precondition of compliance rather than a named duty. Practically: literacy is role-proportionate: an operator reviewing model flags needs different depth than a data scientist or a QP.
Why AI literacy matters now
Employees adopt AI faster than leaders realize. McKinsey finds usage at roughly three times the rate executives assume. In pharma, that gap shows up as shadow tools, unreviewed model outputs in batch records, and audit questions nobody prepared for. AI literacy is the organizational skill to use, review, and govern AI proportionate to role and risk, not a data-science degree for every operator.
In pharma practice
Literacy looks different by role: an operator needs to know when a vision-system flag is off-label input; a validation lead needs context-of-use and independent test data; a QP needs traceability from model output to human verification. The EU codified this in Article 4 (Feb 2025); the US expects competent personnel through FDA credibility and Part 11 review even without a named statute.
Frequently asked questions
Is AI literacy the same as AI training?
Training is one evidence type; literacy is the outcome: judgment to deploy, challenge, and document AI use correctly. Article 4 expects measures matched to role, risk, and context of use, not a one-size e-learning checkbox.
Who needs AI literacy in a pharma company?
Anyone who touches AI outputs that influence GMP decisions, quality records, or submissions, from shop floor review of vision flags to validation leads qualifying ML systems and QA staff releasing agent-assisted drafts. Depth scales with decision consequence.
Which GMP areas need AI literacy first?
Manufacturing (vision inspection, sensors, predictive maintenance), QA/QC (deviations, CAPA, batch record review, trending), validation and IT (qualified systems and agent harnesses), document control and training (RAG search and drafting aids), and release-adjacent roles where human judgment is statutory. The literacy standard is the same; the evidence depth follows context of use.
Related terms
Benchmark your AI literacy
Free 20-question assessment, eight modules, and instant scoring, built for pharma quality and manufacturing teams.
Open the Literacy CenterBrowse all terms
See the full 71-term glossary index or search interactively in the Literacy Center.
Back to glossary hub →Educational content only, not legal or regulatory advice. Regulatory guidance cited here includes drafts (FDA AI credibility guidance; EU GMP Annex 22) as of August 2026; verify against final texts before relying on them in submissions. Company-reported figures (Merck CSR timings, Sanofi results) are labeled where used. MIT's ~95% pilot figure carries its own caveat (~150 interviews, contested definitions, not peer-reviewed). Re-check sources on module finalization.