AI guidance in clinical development: what industry is asking for

The European medicines regulatory network is preparing AI guidance in clinical development. At the HMA-EMA group focused on AI with industry stakeholders, which met virtually on 5 October 2026, regulators said a concept paper is being finalised and public consultation will follow.

Industry input on AI guidance in clinical development

Industry is drafting a technical feedback document to show regulators where clarification would help most. The areas named so far are risk categorisation and proportionality, validation and evidence requirements, third-party AI tools, documentation and transparency, lifecycle oversight of adaptive AI systems, and regulatory submission expectations. Industry will also flag areas where AI-specific guidance may not be needed. Regulators encouraged submissions before or during the consultation.

A horizontal guidance on general machine-learning principles is progressing in parallel, and the two are meant to complement each other. The group’s next meeting is scheduled for 2 February 2027.

Why it matters

For health technology assessment teams, the point to watch is the evidence bar. The group noted that evidentiary expectations may differ with the intended regulatory purpose, and that terms such as synthetic data and digital twins are still read differently across regulators. The network frames the work across medicines, clinical trials and medical devices. The annual multi-stakeholder data forum and AI workshop on 12-13 November 2026 is the next scheduled occasion to follow the topic.

Source: HMA-EMA group focused on AI with industry stakeholders, summary notes (EMA/227688/2026)