AI in Safety‑Critical Design: Where It Belongs and Where It Does Not

Thursday, 11 June 2026

Artificial Intelligence is rapidly becoming embedded across engineering tools, conference agendas, and industry narratives. “AI‑enabled” is increasingly positioned as a default indicator of progress and competitiveness. In many domains, that momentum carries limited risk. In safety‑critical design environments, it does not.

Structured engineering methods such as FMEA exist to enforce discipline, traceability, and accountable decision‑making where failure has operational, regulatory, or human consequences. This session examines how AI should be positioned within those environments—not as a replacement for structured practice, but as a supporting capability with clearly defined boundaries. Rather than focusing on automation or novelty, the discussion explores how AI can assist analysis, consistency, and knowledge management while preserving human judgement and methodological rigour.

Attendees will gain practical guidance on integrating AI responsibly into safety‑critical workflows, ensuring technological progress strengthens engineering integrity rather than diluting it.

Key talking points:

Why safety‑critical design demands different rules for AI adoption

Where AI can effectively support structured methods such as FMEA

Risks of ungoverned or trend‑driven AI use in safety‑critical workflows

Preserving human accountability and decision authority in AI‑assisted design

Practical principles for responsible AI integration in regulated environments