TECHCON 2026

Morning Sessions

As AI accelerates software delivery and empowers adversaries, organisations face a critical inflection point. Join Google Threat Intelligence VP Sandra Joyce as she explores the three core challenges shaping security today: protecting code generated at unprecedented velocity, safeguarding high-value AI workloads from theft and abuse, and withstanding automated adversaries operating faster than any human team can react.

This session will unpack why traditional approaches are unsustainable against autonomous threats, before outlining how defenders can evolve their security posture to regain the advantage.

Adversaries are no longer just adopting new tools—they are operationalising AI to move faster, strike harder, and evade detection with unprecedented sophistication. In this session, Adam Meyers, SVP of Counter Adversary Operations at CrowdStrike, traces the evolution of AI-enabled adversary tradecraft—from early automation to ML-assisted reconnaissance, through today's GenAI-powered operations running at machine speed.

Drawing on real-world intelligence, Adam will examine how threat actors leverage LLMs for hyper-personalised phishing, AI voice synthesis for identity deception, and automated tools to accelerate vulnerability exploitation—illustrated through CrowdStrike's tracking of Evasive Adversaries.

The session closes with a focused look at Singapore's evolving threat landscape—its unique exposure as a regional financial and diplomatic hub—and the actionable defensive strategies organisations must adopt to stay ahead of the AI-enabled adversary.

Afternoon Sessions

Since the start of 2026, AI agents have completely reshaped the landscape of vulnerability discovery and sent defenders scrambling to patch an unprecedented wave of AI-discovered vulnerabilities before attackers—often running their own AI agents—could exploit them. This talk will trace how the frontier models' cybersecurity capabilities grew from "barely usable" in mid-2024 to "nearly superhuman" today, from our perspective building AI agents for web penetration testing as well as our newer work finding vulnerabilities in widely deployed, hardened software like Google Chrome. Brendan will conclude by sketching how he thinks offence/defence balance will play out as we go deeper into the ongoing "vulnpocalypse".

The release of frontier AI models has fundamentally altered the cyber battlefield, shifting the advantage toward those who can execute at machine speed. While defenders look to AI to streamline operations, threat actors are weaponizing these same large-scale models to dismantle traditional security perimeters.

In this session, Unit 42 experts lift the hood on the next generation of automated cyber warfare. Sharing exclusive, real-world red team findings from hands-on testing with Mythos, this presentation reveals exactly what happens when frontier AI is unleashed in an offensive context. We will demonstrate how advanced models provide a dangerous step-function improvement for attackers, drastically lowering the barrier to entry while scaling highly sophisticated, multi-stage attacks.

Attendees will get an inside look at how Mythos like models dynamically authors zero-day exploits, adapts to target environments on the fly, and systematically bypasses legacy security controls that rely on static signatures and human-speed response.

Beyond the threat landscape, this session is a practical blueprint for resilience. You will walk away with actionable, field-tested strategies to modernize your defence and neutralize AI-driven threats before they hit your production environment.

Cybersecurity has traditionally focused on detecting and responding to attacks after adversaries have already taken action. This session explores a more proactive approach: continuously identifying, disrupting, and increasing the operational cost of cybercriminal infrastructure before attacks succeed. Attendees will learn practical strategies for making it harder, slower, and more expensive for threat actors to operate, and why persistent disruption can be one of the most effective forms of cyber defence.

We will share insights on how black-box exploitation can be pivoted into white-box analysis to uncover deeper and more impactful vulnerabilities, as well as how manual and AI-assisted workflows can work hand in hand to accelerate offence-informed defences. We will illustrate these insights drawn from our experiences in pen-testing web applications and security appliances.

Frequently Asked Questions