The insurance industry is moving from a defensive approach to AI exposure, using existing policies, restrictive wording, and selective exclusions – towards affirmative and purpose-built coverage.

However, major questions remain unresolved.

• What must insurers understand before AI risk can be underwritten with confidence?
• Which failures can genuinely be prevented?
• What level of governance, auditability, and human oversight is required?
• Who should build the shared incident data and risk intelligence infrastructure needed to make AI insurable at scale?

This session will bring together The Spark’s prevention and execution thesis, Aon’s strategic market lens and QBE’s practical carrier perspective, to explore what a credible AI-risk insurance model could look like, and what the industry must do now to build it.