Chicago Cybersecurity Conference

AI isn’t conjuring new vulnerabilities out of thin air — but it is dramatically accelerating every stage of the attack chain, from reconnaissance and initial exploitation through to lateral movement and data exfiltration. It’s also making user-targeted attacks like phishing far more convincing and scalable. Security has never really been an onion; it’s a stack of Swiss cheese, and the goal is to make sure the holes don’t line up. AI is making it easier than ever for attackers to find the gaps before you do. The answer isn’t a new tool or a new framework — it’s executing the fundamentals with a level of rigour the industry has rarely demanded before. 90% isn’t good enough anymore. And here’s the uncomfortable truth: your existing tools can’t tell you that you’re only at 90%. No agent-based tool can grade its own homework — it can’t tell you where it isn’t deployed, where it’s broken, or where its configuration has drifted from policy. This talk makes the case for why real-time, cross-platform asset intelligence is the foundation on which brilliant basics must be built.

AI is generating code faster than static tools can evaluate it. The attack surface is growing, but your security stack is still guessing. This session makes the case for runtime as the missing layer — instrumentation that verifies what actually executes, detects attacks at the point of exploit, and connects pre-production findings to production threats through a single intelligence graph. Walk away with a concrete framework for security that works at the speed AI ships code.

Most security programs measure effort — not outcomes. Organizations patch thousands of vulnerabilities, deploy dozens of tools, and run annual tabletop exercises… but when an attacker shows up, none of that matters.

What matters is whether they can prove their defenses actually work.

In this talk, Horizon3.ai shares how leading organizations are using autonomous pentesting to see their environment through the attacker’s eyes — continuously, safely, and at scale. By shifting from assumptions to proof, they’ve learned to:
• Prioritize what’s exploitable. Focus limited resources on the weaknesses that truly put the business at risk that are known to be abused by threat actors.
• Quickly fix what matters. Close the loop from find → fix → verify and reduce your exploitable attack surface.
• Reduce attacker dwell time. Use pentest results to precisely deploy honeyTokens to detect compromise early, and to continuously prove your EDR and SIEM are tuned and working as intended.

Cyber resilience isn’t about being perfect — it’s about getting better over time. And the only perspective that truly matters is the attacker’s.

For two years, agentic AI in the SOC has been a prediction. In 2026, it shipped. This session is a field report from inside the shift. We’ll start with what’s actually changed in the threat landscape — data extortion incidents grew 11x year over year, attackers exploit edge-device vulnerabilities in a median of zero days, and 65% of intrusions now come through legitimate remote access tools. Then we’ll look at what AI is delivering in security operations today: 80-day faster breach lifecycles, $1.9 million saved per breach, and case resolution measured in minutes instead of hours. We’ll cover the new attack surface agentic AI creates — memory poisoning, tool misuse, prompt injection — and the governance frameworks that make autonomous defense trustworthy. Attendees will leave with a clear view of where AI delivers real value now, what’s coming, and concrete steps to take back to their organization.

A good fake Rolex uses Swiss movement, sapphire crystal, 904L steel — every component genuine. Only the crown and the intent are fake. That’s phishing in 2026: real SendGrid accounts, real Cloudflare CAPTCHAs, real Google redirects. The only counterfeit is the intent. Josh Bass will break down what 5,000 real SEG-bypassing attacks reveal about why pattern-matching defenses fail silently and what security leaders need to evaluate differently.

AI coding assistants promise massive productivity gains, but recent academic research shows they also introduce instability, security weaknesses, and developer overconfidence. In this talk, I present data from large-scale studies demonstrating how AI-generated code affects delivery performance, vulnerability rates, and developer decision-making. I then outline a practical, end-to-end approach for reducing AI-induced risk while preserving the benefits of AI-assisted development.

Artificial Intelligence has revolutionized how we work, but it has also armed cybercriminals with unprecedented capabilities. From automated reconnaissance to “Clickfix” social engineering tactics that bypass reputation-based security tools—including AV, EDR/EPP, firewalls, and SWGs/CASBs—attackers are moving faster and smarter. Are your defenses ready for 2026?

Torq surveyed 450 SOC leaders globally to find out what AI is actually doing inside the SOC. The findings challenged some assumptions — and confirmed others.

We’ll dig into insights including:

Why 97% of security leaders are confident AI can handle triage, but only 35% are using it there
What’s behind the trust barrier that 92% of leaders say is holding AI back
What 85% of leaders mean when they say they want a “unified platform”
Where teams plan to expand AI over the next 12 months — and what’s standing in their way

AI has evolved from a research concept into a transformative force reshaping how businesses operate and defend against cyberthreats. Its rapid growth brings powerful benefits — automating routine tasks, improving efficiency, and enabling faster, more accurate threat detection. But AI is a double-edged sword. While security teams use it to predict and stop attacks, cybercriminals exploit it to craft sophisticated phishing campaigns, deepfakes, and adaptive malware.

This session explores AI’s evolution, its rapid rise in business environments, and its expanding role in today’s cybersecurity arms race. You’ll walk away with practical insights and clear steps to help your organization use AI securely, responsibly, and effectively.