Clinical Scorecard: AI in Drug Development: Moving Beyond Early Use Cases

At a Glance

Key Highlights

• EMA and FDA published joint guiding principles for good AI practice.
• Focus on context of use and risk-based controls for AI applications.
• Emphasis on lifecycle management and accountability in AI outputs.
• AI should be embedded in regulated workflows for better governance.
• Regulatory expectations are evolving to ensure trustworthy execution.

Guideline-Based Recommendations

Diagnosis

• Evaluate AI in the context of its intended use and influence.

Management

• Implement clear accountability and human oversight in AI applications.

Monitoring & Follow-up

• Establish defined monitoring and drift triggers for AI outputs.

Risks

• Manage change control proportional to risk and ensure documentation.

Patient & Prescribing Data

Not specified; applicable to organizations in drug development.

Focus on embedding AI within regulated workflows to enhance oversight.

Clinical Best Practices

• Maintain version control that is audit-relevant for AI components.
• Ensure clear documentation and review processes for AI changes.
• Integrate AI deeply within operational systems for better governance.

Related Resources & Content

• EMA-FDA Joint Principles
• EU AI Act Timeline
• FDA Guidance on Electronic Records