Navigating the future of Dual-Use AI at DEFCON 34
Announcing the AI Village keynote panel at DEFCON 34: “The Defender’s Dilemma: Releasing Dual-Use AI Models as Capabilities Climb”
As artificial intelligence capabilities accelerate, cybersecurity is entering a brand-new era. The line between defensive and offensive technology is blurring faster than ever. Powerful AI systems are now dual-use by nature, offering incredible capabilities to secure our networks, while simultaneously handing adversaries new tools to exploit them.
This Friday at DEFCON 34, we are bringing this critical conversation to the main stage.
This year’s AI Village Keynote panel, titled “The Defender’s Dilemma: Releasing Dual-Use AI Models as Capabilities Climb,” will explore the practical challenges that developers and security labs face today. As these frontier models grow stronger, the industry must answer a pivotal question: how do we ensure that defenders stay ahead?
We have assembled an incredible group of industry leaders to dive into these issues. The discussion will feature panelists Heather Adkins (VP of Security Engineering at Google), Jason Clinton (Deputy CISO at Anthropic), and legendary cryptographer and author Bruce Schneier, and will be moderated by AI Village Co-Chair Emanuel Gawrieh (Senior Chaos Containment Engineer at Google).
Together, the speakers will share firsthand insights into the way teams evaluate a model’s true capabilities before it ever reaches the public. They will weigh the distinct security trade-offs of different release strategies, the evolution of zero trust in the agentic era, and look ahead at what the next wave of exponential AI growth means for the future of global digital defense.
Whether you are a hacker, defender, security researcher, or simply curious about the future of Dual-Use in AI, this is a conversation you won’t want to miss.
Session Details
- Date: Friday, August 7, 2026
- Time: 1:30 PM PDT
- Location: Las Vegas Convention Center (West Hall), Mainstage 1
We hope to see you there!