Main Program

A case study shows how one enterprise consolidated fragmented AI guidance into a governable control framework covering RAG, copilots and agentic systems.

A Think Tank on cyber security as part of operational resilience for transport, health, water and utilities, and on lessons from regulated industries.

Thursday afternoon at 3.50pm, Ask an Expert 1 turns to leadership, governance and accountability with Theo Nassiokas of Wesfarmers Limited.

Building and running a crisis management exercise from scratch gives participants a repeatable model they can take back to their own organisation.

An open-source spooling architecture stores large MCP tool responses outside the context window, returning only the summary an LLM actually needs to work with.

A hands-on SANS FOR508 workshop: direct an agentic AI through a simulated DFIR investigation, then approve or reject each finding before it enters the case.

The board wants to know how long until we are operational again, and the session's answer is that standard security metrics do not predict it.

Tane Hunter built machine learning models at Australian research hospitals and wrote Full Stack Human. He has given over 500 keynotes across five continents.

A panel on the new Cyber5050 report: what the survey data says about cyber's skills shortage and the barriers limiting women's participation and progression.

Attackers now go after the accounts an organisation relies on and AI systems acting on its behalf, and this session shows how layered defence has to change.

Rezza Moeini draws on more than 300 hiring decisions to explain where candidates lose him, and why he now hires for judgement as entry-level work is automated.

As AI agents start shopping and paying on people's behalf, this session maps the new commerce protocols and controls needed to secure transactions.

As AI agents gain power to manage cloud infrastructure, Tier 0 must expand beyond domain controllers to cover identity platforms and agentic access.

A randomised trial found mandatory phishing training makes employees more likely to click, and this session explains why the model is broken and what works.

Perspectives from governance, risk and security operations show how credibility is built, challenged and unevenly recognised for women in cyber security.

Generative AI embedded across enterprise tools quietly accumulates permissions and authority, and this session proposes governing it as a non-human insider.

Q-Day will break RSA and ECC, and adversaries are already harvesting encrypted data to decrypt later. What building a sovereign quantum capability would take.

Most AI threat intelligence tools produce impressive but unusable summaries, and this session shows how to build systems that produce structured intelligence.

Australians lost $2.18 billion to scams in 2025. How to read the Scams Prevention Framework as an operational blueprint that turns reporting into disruption.

Understanding attacker psychology and system environments in critical infrastructure helps organisations cut compromise risk when essential services are hit.

A live ransomware simulation shows exactly where layered security controls detect and contain an attack, and how recovery planning speeds up resumption.

Recent regulatory and legal decisions reveal what counts as reasonable, adequate cyber risk measures, giving organisations a framework to test readiness.

A pattern-recognition method for reading a security fix to predict related vulnerabilities and future disclosures from the same vendor.

Wireless and OT red teaming still depends on human intuition and specialist hardware, revealing weaknesses automated and AI-driven testing consistently miss.