When Seeing Isn't Believing: Combatting Deepfakes with iProov's Dominic Forrest Video

Can cyber defenses built around human attackers keep pace with AI? Brad Medairy, president of Booz Allen's National Cyber Practice and a McCrary Institute senior fellow, joins Frank Cilluffo to discuss what his team's testing reveals about AI's ability to conduct cyber operations and navigate industrial systems. They examine how these capabilities lower barriers for attackers, why deploying defensive AI requires trust and human accountability, and what executives should do to assess readiness. The conversation also explores active cyber defense and the need to keep essential services running through a compromise. Main topics AI-driven cyber operations Frontier model testing U.S.–China capability gaps Industrial systems at risk AI attack simulations Human oversight and accountability Deception and active defense Operating through compromise Board oversight and readiness Key quotes "Today, with an agentic attack, with an AI-enabled cyber operator that can operate at speed and scale, I think the defender does not have the advantage. It can't move at that pace." — Brad Medairy The way that we've seen models be able to rip apart these [OT] environments, understand the protocols, understand the control system, it's really phenomenal in terms of the capabilities that they provide… the risk to critical infrastructure is at an all-time high." — Brad Medairy "I don't think that we're going to defeat the threat of the future that's so transformative by just automating past operating models and approaches." — Brad Medairy "These adversary capabilities are progressing much faster than cyber defenses. And so you're gonna have to continually assess your environment, continually assess your controls, continually evolve your program." — Brad Medairy "We're inventing new operating models, we're inventing new technologies for cybersecurity. And I think now is our time to really define the future. It's going to take… really working this problem together." — Brad Medairy Resources discussed Booz Allen's Cyber Weapons Index McCrary Institute–Booz Allen joint paper: "AI-enabled threats demand mission-speed resilience – here's how to build it" McCrary Institute–Booz Allen joint paper: "Outpacing AI-enabled cyber threats: A playbook for government CISOs" About the Guest: Brad Medairy is president of Booz Allen's national cyber business, leading the development of defensive and offensive cyber solutions to address America's toughest cyber challenges. His work spans AI, machine learning, cloud innovation, and intelligence tradecraft. He leads thousands of cyber operators, security researchers, and software engineers who secure supply chains, protect critical infrastructure, and defend 5G networks, space systems, and advanced weapons platforms. His team counters sophisticated adversaries for customers including the intelligence community, U.S. Cyber Command, FBI, CISA, and the Department of War.

Keeping power, water, and other essential services running requires planning for what happens when cyber defenses fail. On this episode of Cyber Focus, Frank Cilluffo speaks with Stephanie Crowe, head of the Australian Cyber Security Centre, and Nick Andersen, acting director of the Cybersecurity and Infrastructure Security Agency (CISA), about how two close allies are preparing critical infrastructure to operate through compromise. The conversation examines how governments and industry identify their most critical systems, turn shared intelligence into joint action, and prepare for attacks that cross sectors and borders. Crowe and Andersen also discuss how AI is changing cyber defense and what that means for infrastructure investment, workforce skills, and decisions in the boardroom. Main Topics Operating through cyberattacks Protecting critical assets and dependencies China's infrastructure prepositioning CI Fortify and joint industry planning U.S.-Australia and Five Eyes cooperation AI defense and aging infrastructure AI's energy demands Workforce skills and boardroom decisions Key Quotes "You have to be able to fight through compromise now, which is a really different way of thinking about the cyber threat." — Stephanie Crowe "We've come a long way from thinking about our number one cyber risk being data breaches and theft of information into a world where service disruption is now the thing that we have to really consider and plan in our cyber risk management." — Stephanie Crowe "And just to be completely honest, Australia really bailed us out when we were in the middle of our shutdown… Australia led from the front." — Nick Andersen "If the Congress decided that in the next budget that CISA needed to be quintupled in size, we're still not going to have enough resources to be able to meet every single area that we'd like to be able to stray into with our attention." — Nick Andersen "We know that we're seeing exceedingly aggressive, well-orchestrated campaigns, in particular from entities like China, that are engaged in a significant way to preposition across our critical infrastructure to hold it at risk at a time and place of their choosing." — Nick Andersen Links and Resources Cybersecurity and Infrastructure Security Agency (CISA) Australian Cyber Security Centre: Guidance and Resources About the Guests: Stephanie Crowe is Deputy Director-General of the Australian Signals Directorate and Head of the Australian Cyber Security Centre. She has held senior roles in signals intelligence and cybersecurity, leading efforts to strengthen Australia's cyber defenses, build threat intelligence capabilities, and support industry during major cyber incidents. Nick Andersen is Acting Director of the Cybersecurity and Infrastructure Security Agency (CISA), where he leads efforts to reduce risks to America's cyber and physical infrastructure. A U.S. Marine Corps veteran, he has held senior cybersecurity roles at the Department of Energy, the White House, and in the private sector.

As AI expands the capabilities of cyber adversaries, does the federal government have the workforce and tools to keep pace? Rep. James Walkinshaw of Virginia joins Frank Cilluffo to discuss cuts at the Cybersecurity and Infrastructure Security Agency (CISA), their implications for critical infrastructure, and his push for an independent assessment of the agency's capacity. They also examine responsible AI adoption in government, the risks of unauthorized AI tools, and the public anxieties driving opposition to data centers. Walkinshaw outlines efforts to strengthen the federal technology workforce, streamline security approvals, and improve accountability for basic cybersecurity practices. Main Topics CISA staffing and readiness Local cybersecurity support Federal cyber workforce Responsible AI adoption Shadow AI risks Data centers and public trust FedRAMP reform CLEAR Act and event logging Bipartisan cyber leadership Key Quotes "At the precise moment when we most needed CISA to be operating at peak efficiency, it was systematically dismantled." — Rep. James Walkinshaw "When it comes to AI especially, cybersecurity can't be an afterthought. It has to be built in from the very beginning." — Rep. James Walkinshaw "People are starting to see data centers as the physical embodiment of the anxieties we all have around AI." — Rep. James Walkinshaw "If you want to bring progress in AI to a screeching halt overnight, that will happen if something goes really, really bad. If you don't get the safety piece right." — Rep. James Walkinshaw "Federal agencies currently are not even meeting the requirements in terms of event logging. They certainly, for the most part, couldn't answer any of those questions about time to detect or time to mitigate. So our CLEAR Act will require some reporting to Congress around event logging." — Rep. James Walkinshaw Links and Resources The CLEAR Act: Walkinshaw and Bacon's announcement FedRAMP CISA: State and Local Cybersecurity Grant Program updates About the Guest: Rep. James Walkinshaw represents Virginia's 11th Congressional District and serves on the House Homeland Security and Oversight and Government Reform committees. Before entering Congress, he served on the Fairfax County Board of Supervisors and spent more than a decade as chief of staff to Rep. Gerry Connolly. He is also founder and co-chair of the Congressional Federal Workforce Caucus.

Can the United States compete with China in AI while managing the risks the technology creates? Host Frank Cilluffo welcomes McCrary Institute senior fellows Bill Evanina and Matt Hayden for a discussion of the difficult choices behind that question. Evanina, who previously led U.S. counterintelligence, and Hayden, a former senior CISA official now at GDIT, assess expectations for the Trump–Xi summit, the limits of cyber cooperation, and the trade-offs behind chip export controls. The conversation also examines Chinese influence efforts surrounding data centers, the importance of energy to AI competition, the stakes of quantum computing, and whether familiar approaches to accountability can work when an AI agent becomes an insider threat. Main topics Trump–Xi summit expectations Volt Typhoon and cyber deterrence Chip exports and technology theft S.–China crisis communications Chinese influence on data center debates AI, energy, and the quantum race AI safeguards and accountability Taiwan and semiconductor dependence Key quotes "We cannot be AI dominant unless we're energy dominant. We can't be. And when you look at AI, it's not AI in its own, it's that it accelerates everything else." ­­— Frank Cilluffo "And so when we look at AI, it's not a matter of having the most advanced models. That helps. But it is a matter of getting that implementation into every corner of industry so that those efficiencies are known and used." — Matt Hayden "The big question is AI— should it be considered an insider threat? And the answer was a resounding yes. But what do we do with insider threats? We try to catch them and prosecute them. Are we going to prosecute an AI agent?" —Bill Evanina "From a pure liability standpoint of who we can blame, there's no AI agent that operates without compute. And so somebody's got to empower this thing at an original sin level." — Matt Hayden "As a metaphor, AI is a supermarket gas station on the highway to quantum. ... Whoever wins quantum, they say wins the world. I think you have to win AI before you win quantum." — Bill Evanina Links and resources CISA: Volt Typhoon and persistent access to U.S. critical infrastructure

The federal government wants private companies to play a more active role in cyber operations. A new White House memo would allow vetted firms to support government action against transnational criminal groups. The plan marks a shift from sharing threat data to disrupting threats together. But key questions remain: Who controls the mission? Who carries the risk? Where does private support end and government authority begin? Stacy O'Mara joins Frank Cilluffo to discuss those questions. They examine oversight, liability, AI and critical infrastructure. O'Mara also draws lessons from SolarWinds and the war in Ukraine. Her bottom line: Trust must exist before a crisis, and new programs must be tested before they are scaled. Main Topics Covered New rules for private cyber operations From information sharing to joint action Authority, oversight and private-sector risk AI, critical infrastructure and human judgment Trust and readiness before a crisis Key Quotes "We can put all these partnerships in place, create a 70-page memo on how to do it. But at the end of the day, if there's a significant event, it's going to be the leadership of major ISPs, telecoms, and cybersecurity providers who are picking up the phone and calling each other in the middle of the night." — Stacy O'Mara "My gut tells me it is going to take a serious incident, and it would probably include loss of life. …But until we actually go through it and suffer some of the consequences, I don't know that we'll actually be able to put everything into place" — Stacy O'Mara "It can't just be: 'We're going to collaborate, and we're going to do all these things together because we're scared and we don't know what's coming.' But it needs to be: 'What are our desired outcomes?'" — Stacy O'Mara "You want AI to perform operations—or activities, rather—at speed and scale. But the human aspect should be around judgment and consequences. That should not be left up to AI." — Stacy O'Mara "Pilot these programs. Get a win. Figure out what works and doesn't work. And then replicate it and scale it. We don't have to do everything all at once." — Stacy O'Mara Relevant Links and Resources White House memorandum: Expanding Capabilities to Combat Transnational Cyber-Enabled Crime CISA Gold Eagle initiative Armadin Guest Bio Stacy O'Mara leads government affairs at Armadin. She previously held cybersecurity policy and government affairs roles at FireEye, Mandiant and Google. She later advised technology companies at Venable.