Most organisations treat AI risk as “just cyber”—but the targets, vectors, data classes, and liabilities are materially more extensive and significant. We’ll examine AI‑specific threats (adversarial inputs, model inversion, extraction, poisoning, prompt injection), how these map to real legal exposure (contract, negligence, IP, defamation, breach of confidence), and why existing policies may respond even without an explicit “AI” endorsement. We’ll also translate the fast‑moving rulebook—DORA, NIS2, the UK Cyber Security and Resilience (NIS) Bill, and the EU AI Act—into practical governance and control expectations for boards, CISOs, underwriters, and brokers. You’ll leave with a crisp view of controls that actually matter for AI systems (beyond GDPR), covering:

• Implementation & Ops 
  How AI systems are actually deployed and stabilised in enterprises, and how that changes the security control set (beyond classic network controls).
• Cyber Controls in AI Environments
  Identity, monitoring, attack‑surface management for AI, and controls for model abuse, prompt injection, extraction, and poisoning.
• Governance, Legal Risk & Insurance
  Existing law already catching AI behaviour; DORA/NIS2/UK NIS Bill/EU AI Act; contractual allocation; “silent AI” cover; wording evolution.

Capped at 20 delegates this roundtable has been tailored to the needs of senior business; IT-security; audit; legal and compliance professionals and underwriters and brokers who work in the Lloyd's and London insurance market.  If this roundtable is relevant to colleagues within your company, then please give notice about this event. Thank you.