AI Fluency as a Compliance Control

Most GMP organizations have policies for AI, but policy doesn't control what happens at the bench, in the deviation system, or during a batch record review. This session reframes AI fluency as a measurable, auditable compliance control and teaches a practical method (ALIGN) for converting common GenAI failure modes into workflow-level controls, including a GxP use-case classification framework, intended use boundaries, and a minimum SOP architecture.

The compliance exposure lives in the last mile: how individuals actually use AI day to day. A well-vetted tool used without structure creates the same risk as an unapproved one, and most organizations haven't closed that gap yet.

This session introduces ALIGN, a five-step operational method (Aim, Learn, Input, Guard, Notice) that maps each known GenAI failure mode to a specific workflow control. We will cover:

• Failure modes and the case for operational controls.** Why governance frameworks and vendor qualifications are necessary but not sufficient, and where compliance risk actually shows up in daily AI-assisted work.
• The ALIGN method applied to GMP workflows.** How to define intended use for AI-assisted tasks, understand tool capabilities and constraints, standardize prompts with evidence-first drafting patterns, enforce data boundaries, and build verification checkpoints scaled to risk.
• GxP use-case classification.** A three-tier framework (Non-GxP, GxP-supporting, GxP-critical) that right-sizes validation effort, review intensity, and documentation requirements. We'll walk through how this classification maps to ALIGN's risk-tiering approach and helps teams avoid both over-control and under-control.
• Intended use boundaries.** How to write clear intended use statements for AI-assisted tasks that prevent "tool creep," define where human accountability begins, and make the scope of AI involvement auditable.
• Onboarding AI into GMP environments.** A phased approach to bringing AI-assisted workflows into production, from initial use-case triage and tool assessment through controlled pilots with verification checkpoints to go-live readiness criteria and ongoing monitoring.
• Inspection-ready governance.** The minimum SOP architecture needed to support AI-assisted work under GMP, covering use-case registration, change control for AI configurations, model and version tracking, data boundary enforcement, and the AI assist record as a documentation backbone.

The session includes a worked example in regulated content drafting that shows how ALIGN controls caught a plausible but incorrect AI output before it reached a review gate. We'll close with a minimum viable control package of practical artifacts that connect daily practice to governance frameworks, including NIST AI RMF, ISO/IEC 42001, and GxP principles.

• QA/QC, Manufacturing, and CMC Leaders
• Quality Systems and CSV/CSA Specialists
• Regulatory Operations and Compliance Professionals
• Anyone responsible for how AI tools are used in GMP environments

Date: Thursday, June 04, 2026

Doors Open: 5:30 pm

Location: Condon Hall (on the UW Campus), 1100 NE Campus Prkwy, Seattle, WA 98105

Don't miss this unique chance to connect with peers and expand your knowledge. Space is limits so register now to secure your spot!

Alexa (Lex) Kopf

Founder, LabScale AI

https://www.linkedin.com/in/alexa-kopf/

https://labscaleai.com/

Parking is available for the building (W41) at the end of NE 41st ST on the North side of the building. Parking in this lot is paid until 6:00 pm.

There is paid street parking as well.

Know what’s Happening Next — before everyone else does.

Get App

Host or Publisher PDA Pacific Northwest Chapter