AI insurance in practice: an interactive workshop on risk, assurance and underwriting

About this workshop

In February 2026, techUK convened an expert panel on AI insurance, assurance and risk mitigation, surfacing a clear message: AI insurance is emerging from an immature market, where many risks remain difficult to quantify, and where the lessons of cyber insurance (both its successes and its failures) are directly relevant to how this market develops.

This workshop builds on that discussion, and on techUK's related work on AI procurement and Investors for AI Assurance, to move the conversation from diagnosis to practical detail. Rather than a panel discussion, this session is designed as an interactive workshop: small groups of practitioners, insurers, assurance providers and policymakers will work through real scenarios and open questions to generate the insights that will feed directly into a techU’s brief on AI insurance later this year.

Why this matters now

The AI insurance market is moving quickly, but foundational questions remain unresolved: how to price risk that is probabilistic rather than deterministic; how to handle correlated exposure when many organisations rely on the same foundation models; who is liable when an AI system causes harm; and how insurance, assurance and regulation can work together rather than at cross purposes. Getting this right matters not just for insurers and the assured, but for the wider AI ecosystem, informing everything from how organisations procure AI to how investors assess the risk profile of the companies they back.

What we'll cover

Participants will work through facilitated breakout sessions on:

• Quantifying the unquantifiable: what would workable underwriting criteria for AI systems actually look like, and what evidence should insurers ask for?

• Standards and evaluation: what would a common assessment framework for AI systems need to include to be credible for both assurance and underwriting purposes?

• Systemic and correlated risk: how should the market think about concentration risk from shared foundation models and third-party AI dependencies?

• Incident reporting: what would a workable AI incident reporting regime look like, and who should hold the data?

• Liability and causation: working through real-world scenarios to test where responsibility sits across the AI supply chain.

The session is structured to generate concrete, citable input for the whitepaper practitioners' views, points of consensus, and the genuine areas of disagreement that still need to be worked through.

Who should attend

This workshop is designed for a mixed room, including:

• Insurers and underwriters developing AI-specific products

• AI assurance and evaluation providers

• Legal, risk and compliance leads working on AI liability

• Investors assessing AI risk in portfolio companies

• Policymakers and regulators engaged in AI governance

• Organisations deploying AI at scale who want a voice in shaping this market

Outputs

Insights from this workshop will directly inform a brief paper on AI insurance.

Background reading

• AI Insurance, AI Assurance and Risk Mitigation: Key Insights from techUK's Expert Panel

• Understanding AI Insurance and the Role It Plays in AI Assurance