CMC Insights
November 2026

Monday, November 9, 2026

FDA’s Evolving Approach to AI in Pharmaceutical Development and Manufacturing

Verta Life Sciences

Understanding Regulatory Expectations for AI in Drug Development, CMC, Quality, and Manufacturing

Artificial intelligence is transforming pharmaceutical development, from drug discovery and clinical research to manufacturing optimization, quality management, and regulatory submissions. As adoption accelerates, the FDA is establishing a clearer regulatory approach to ensure that AI-generated evidence and AI-supported decisions remain scientifically reliable and appropriately controlled.

Recent FDA guidance, internationally harmonized principles, and enforcement activity highlight a consistent message: AI can improve efficiency and decision-making, but its application must be justified by its intended use, associated risks, and appropriate regulatory oversight.

This month’s newsletter examines the FDA’s evolving approach to AI and the practical implications for pharmaceutical and biotechnology organizations.

Key Takeaways

  • FDA is taking a risk-based approach to AI: The level of evidence, validation, and oversight should reflect the intended use of an AI model and the consequences of incorrect outputs.
  • AI-generated evidence requires demonstrated credibility: FDA’s draft guidance introduces a framework for assessing the reliability of AI models used to support regulatory decisions involving drug safety, effectiveness, and quality.
  • FDA and EMA are aligning on good AI practices: Ten guiding principles emphasize human oversight, data governance, model performance, and lifecycle management across drug development.
  • AI applications in manufacturing remain subject to cGMP requirements: AI-enabled process monitoring, quality control, and manufacturing documentation must operate within established quality systems.
  • FDA enforcement highlights the risks of overreliance on AI: A 2026 warning letter reinforces that AI-generated manufacturing documentation requires appropriate verification and Quality Unit oversight.

1. FDA Establishes a Risk-Based Framework for AI in Regulatory Decision-Making

January 2025 | Draft Guidance

The FDA issued draft guidance titled Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products.

The guidance proposes a framework for establishing the credibility of AI models when their outputs are used to support regulatory decisions related to drug safety, effectiveness, or quality.

Central to this approach is the Context of Use (COU), which defines the specific purpose of an AI model, how its outputs will be used, and the regulatory decisions those outputs will inform.

The framework emphasizes:

  • Defining the model’s intended context of use.
  • Assessing the risks associated with incorrect model outputs.
  • Establishing appropriate credibility assessment and performance evaluation plans.
  • Documenting model development, data sources, assumptions, and limitations.
  • Implementing risk mitigation measures and appropriate regulatory engagement.

Rather than applying identical requirements to every AI application, FDA proposes that credibility assessments be proportionate to model risk and intended use.

Why It Matters

AI used to support a regulatory submission requires a different level of scrutiny than AI used for exploratory research or routine administrative activities.

For example, an AI model used to predict a critical quality attribute or generate evidence supporting product quality decisions may require a more rigorous credibility assessment than a model used to prioritize early-stage research activities.

Pharmaceutical sponsors should identify where AI-generated evidence influences regulated decisions and establish appropriate documentation, verification, and oversight strategies.

Regulatory status: This document remains draft guidance and contains nonbinding recommendations.

Read FDA’s Draft Guidance

2. FDA and EMA Introduce Ten Guiding Principles for Good AI Practice

January 2026 | Joint FDA–EMA Principles

The FDA and European Medicines Agency (EMA) published ten guiding principles for responsible AI use throughout the drug development lifecycle.

These principles apply broadly to the generation and analysis of evidence across nonclinical research, clinical development, manufacturing, and postmarketing activities.

The principles emphasize:

  1. Human-centric design and appropriate oversight.
  2. Risk-based development and validation.
  3. Compliance with applicable standards and GxP requirements.
  4. Clearly defined contexts of use.
  5. Multidisciplinary scientific and technical expertise.
  6. Robust data governance and documentation.
  7. Appropriate model design and development practices.
  8. Risk-based performance assessment.
  9. Lifecycle management and ongoing monitoring.
  10. Clear communication of model capabilities and limitations.

The principles provide a common foundation for responsible AI use while supporting international regulatory alignment.

Why It Matters

For pharmaceutical companies operating across multiple regulatory jurisdictions, the FDA–EMA principles provide a foundation for establishing consistent AI governance practices.

Companies should consider how AI systems are developed, qualified, monitored, and maintained throughout their operational lifecycle.

This includes addressing data provenance, model performance, changes in underlying data, potential performance drift, and appropriate human oversight.

Regulatory status: These are guiding principles, not new binding regulations. Existing GxP and other applicable regulatory requirements continue to apply.

Read FDA–EMA Guiding Principles

3. The FDA’s underlying expectations

FDA Regulatory Focus | Manufacturing and Quality

The FDA has also examined the application of AI in pharmaceutical manufacturing, including its potential to improve process monitoring, process control, predictive maintenance, and quality assurance.

In its discussion paper, Artificial Intelligence in Drug Manufacturing, FDA identified important regulatory considerations associated with AI-enabled manufacturing technologies.

Areas of focus include:

  • Process monitoring and control: Evaluating the reliability of AI models used to monitor manufacturing processes and support process adjustments.
  • Model validation: Establishing appropriate methods for assessing the performance of AI models operating in cGMP environments.
  • Data integrity: Ensuring that data used to develop and operate AI models are accurate, reliable, and appropriately governed.
  • Change management: Managing model updates, retraining, and performance changes throughout the system lifecycle.
  • Quality oversight: Maintaining appropriate controls and accountability when AI supports manufacturing or product quality decisions.

Why It Matters

AI has the potential to improve manufacturing efficiency, identify process deviations earlier, and support more proactive quality management.

However, implementing AI in regulated manufacturing environments requires careful consideration of process validation, computerized system controls, data integrity, and established quality responsibilities.

For CMC and Quality teams, the priority is to establish a clear understanding of how AI influences manufacturing processes and product quality, and to ensure controls are proportionate to the associated risks.

FDA’s manufacturing discussion paper explores potential regulatory considerations; it does not itself establish new cGMP requirements.

Read FDA’s Discussion Paper: Artificial Intelligence in Drug Manufacturing

4. FDA Warning Letter Highlights the Risks of Overreliance on AI

April 2026 | FDA Enforcement Action

An FDA warning letter issued to Purolea Cosmetics Lab provides a concrete example of the regulatory risks associated with inappropriate reliance on AI in pharmaceutical manufacturing.

During its inspection, FDA found that the company had used AI agents to generate drug product specifications, procedures, and master production and control records without adequately verifying their accuracy and compliance with cGMP requirements.

The agency also documented significant manufacturing and quality deficiencies, including inadequate process validation and insufficient Quality Unit oversight.

FDA emphasized that AI-generated manufacturing documentation must be reviewed and cleared by authorized Quality Unit personnel.

Why It Matters

The warning letter reinforces an important distinction: AI can support regulated activities, but it cannot replace regulatory expertise, human accountability, or established quality controls.

Organizations using AI to generate manufacturing documents, specifications, or quality procedures should ensure:

  • AI-generated outputs undergo appropriate technical and quality review.
  • Qualified personnel retain responsibility for regulated decisions.
  • Procedures, specifications, and manufacturing records meet applicable cGMP requirements.
  • AI adoption is integrated into established quality governance and document-control processes.

The enforcement action demonstrates that reliance on AI does not reduce the manufacturer’s responsibility to maintain compliant manufacturing operations.

Read FDA Warning Letter: Purolea Cosmetics Lab

5. What Pharmaceutical Companies Should Prioritize

Together, FDA’s guidance, collaborative principles, manufacturing discussions, and enforcement activity point toward several practical priorities for organizations adopting AI across pharmaceutical operations.

Establish an AI inventory and risk classification. Identify AI applications across development, manufacturing, Quality, and regulatory activities, distinguishing exploratory tools from systems supporting regulated decisions.

Define the context of use. Clearly document the purpose of each AI application, its intended outputs, and the decisions it supports.

Develop risk-based credibility and validation strategies. Establish appropriate performance measures, verification procedures, acceptance criteria, and documentation based on intended use and potential risks.

Strengthen governance and human oversight. Define responsibilities for technical review, quality approval, system monitoring, and regulatory compliance.

Plan for lifecycle management. Implement controls for model updates, changes in operating conditions, data drift, and periodic performance reassessment.

These priorities represent practical measures informed by FDA’s regulatory direction, rather than a single mandatory AI governance framework applicable to every use.

Verta’s Perspective

Responsible AI Adoption Requires Strong Scientific and Quality Foundations

The FDA’s evolving approach signals an important shift in how pharmaceutical organizations should evaluate and implement AI.

The central question is no longer simply whether AI can improve efficiency, but whether its outputs can be trusted for their intended use and appropriately integrated into regulated processes.

For CMC, Quality, and Regulatory Affairs teams, this means applying established scientific principles and quality management practices to new technologies while understanding where additional credibility assessments and lifecycle controls may be necessary.

At Verta Life Sciences, we help pharmaceutical and biotechnology companies prepare for critical CMC, Quality, manufacturing, and regulatory milestones through CMC strategy and gap assessments, process validation, Quality systems, manufacturing readiness, regulatory support, and inspection preparation.

As AI adoption expands, strong quality systems, scientific justification, and regulatory readiness will remain essential to translating technological innovation into reliable, compliant pharmaceutical operations.

Verta Life Sciences
CMC | Quality | Regulatory Support