Cloud, Agentic AI & The Future of Cyber Defence

Chairpersons welcome and opening remarks.

Automation and interconnected APIs have dramatically widened what defenders need to monitor and protect in the cloud. This session breaks down where the modern cloud attack surface is expanding fastest, and what security teams should be prioritising as identity, APIs and automated workflows become primary targets.

Autonomous AI agents are increasingly being given real access and permissions inside enterprise systems, often faster than governance can keep up. This panel of security leaders examines the risks this creates, from over-provisioned access to unpredictable agent behaviour, and what practical guardrails organisations should be putting in place now.

Quantum computing may still be some years away from breaking current encryption, but data harvested today can be decrypted retroactively once it arrives. This session lays out a practical timeline for quantum-safe migration and explains why organisations holding long-lived, sensitive data can’t afford to wait.

Large language models and autonomous agents introduce entirely new categories of vulnerability that traditional application security wasn’t designed to catch. This session covers practical technical approaches to defending against prompt injection, model manipulation and other emerging exploitation techniques targeting AI systems directly.

Retrofitting security onto AI systems after the fact rarely works well. This conversation explores what it genuinely means to embed security into AI development from day one, covering data handling, model access and testing practices that reduce risk before systems ever reach production.

With the US still without a comprehensive federal AI law, state-level requirements expanding and new federal initiatives putting greater focus on AI security and resilience, organisations face an increasingly fragmented compliance landscape. This session unpacks what’s actually changing, where organisations should expect requirements to converge or diverge, and what security and compliance teams should be prioritising now to keep pace with an evolving AI regulatory environment.

A forward-looking session bringing together predictions on the technologies, threats and strategic shifts most likely to shape cyber defence over the next five years. Expect a mix of grounded near-term forecasting and genuine debate about where the bigger, longer-term uncertainties still lie.

As attackers increasingly deploy AI to automate and scale their operations, defenders are responding in kind. This session explores how AI-powered defensive tools are being deployed to keep pace, and what it takes to make that defensive investment genuinely effective rather than just another layer of noise.

Crypto-agility, the ability to swap out cryptographic algorithms quickly as standards evolve, is becoming an essential architectural property rather than a nice-to-have. This panel discusses what crypto-agile infrastructure actually looks like in practice, and the concrete steps enterprises can take today to prepare.

The window between a cloud misconfiguration occurring and an attacker exploiting it is shrinking fast, driven partly by automation on the attacker’s side. This session examines why that gap keeps closing and shares practical approaches defenders can use to close it further from their own side.

A senior security leader shares an honest, first-hand perspective on how agentic AI is reshaping strategic thinking, budget priorities and team structure. The conversation covers what’s genuinely changing versus what’s still hype, and how leaders are adapting their roadmaps in response.