Main Program

Four years of research into how cybercriminals launder crypto proceeds through mixers and privacy chains, and the countermeasures that can disrupt it.

Intelligence tradecraft used in cyber security can identify technology-facilitated abuse in domestic violence cases that defy conventional detection models.

Safeguarding AI-enabled cyber systems depends on who shapes risk decisions, making neurodiversity and inclusive leadership a genuine resilience factor.

Martin Leung traces AI in his security work from early hallucinated CVEs to agentic tools, ending with five questions senior leaders should ask their teams.

AI features are shipping without clear risk ownership, and this session sets out how ISO 42001 closes the accountability gap teams keep avoiding.

DNS-AID offers a practical, standards-based way to let autonomous AI agents discover and trust each other without handing control to a single private registry.

Research on incidents including Gemini memory poisoning introduces the Digital Insider framework for governing AI agents that cause harm with legitimate access.

Ageing assumptions about trusted registries and pre-release reviews are accelerating supply chain risk, and only verified, signed provenance closes the gap.

Treating privacy and legal requirements as blockers rather than embedding them in user stories quietly builds technical debt that surfaces later.

A new governance and accountability model shifts cyber security ownership onto business leaders, embedding it as a core business management process.

The Bunnings tribunal ruling sets the first Australian benchmark for facial recognition technology, outlining what organisations must do to deploy it lawfully.

At 10.55am on Friday, Nigel Phair, regional director APAC at CREST, joins Ask an Expert 1 for leadership, governance and accountability.

Using the military SMEAC planning framework, this workshop turns Australia's 2030 post-quantum cryptography deadline into a one-page operational order.

A device-free, hands-on workshop uses drawing and collage to build the bias-recognition and evidence-based reasoning skills that AI can't replace.

Defining system boundaries, building a defensible evidence package, and managing the assessor relationship are the keys to a smooth IRAP assessment.

A workshop on why vendor due diligence struggles to assess autonomous AI agents, and how certification applies continuous adversarial stress-testing.

The build-versus-partner decision is costed against the move from MSP to MSSP, including what specialist talent and round-the-clock coverage cost in-house.

Treating AI as one monolithic risk category leads boards to apply generic controls that miss how differently AI systems actually behave, fail and expose data.

Breaches often exploit small, preventable gaps created by limited capacity and unrealistic coverage expectations, not more advanced adversaries.

Zero Trust is the stated comparison, and what it points to is continuous, reprioritisable security delivery across the whole cyber security function.

A multi-agent framework combining graph theory and large language models helps SOC analysts reason about physical dependencies when detecting ICS threats.

Testing the Essential Eight against SolarWinds, XZ Utils and a real malicious package shows it covers barely half of a modern supply chain attack surface.

An airline captain and cyber practitioners apply aviation safety models, including Swiss Cheese and SHELL, to redesign security around real human performance.

Age-verification and AML/CTF rules force mass collection of biometric and identity data. A legal and technical panel asks who is liable for these honeypots.