AI vs. AI: Why the Future of Cybersecurity Means Humans Are Out of the Loop
Today, former NSA Technical Director and Synack co-founder, Mark Kuhr, breaks down the rapid evolution of generative AI, agentic AI, and their deep impact on cybersecurity across government. Mark explains that the proliferation of AI tools, like ChatGPT, Claude, and Gemini, creates urgent new risks for agencies around data control and network security. He stresses the importance of securing sensitive data by encouraging the use of government-developed AI tools instead of consumer-grade models, avoiding potential leaks similar to past shadow IT problems.Drawing on his public and private sector experience, Mark highlights the critical privacy concerns both sectors face as employees blend personal and work-related AI usage, often without proper controls. He discusses how the adoption of government-specific AI tools, like GSA’s new initiative, is a smart move toward safeguarding agency data. Mark also introduces agentic AI—the next frontier where multiple specialized AI agents collaborate autonomously to solve problems, mimicking human teamwork. Mark emphasizes that in a world where adversaries are using AI to exploit vulnerabilities within minutes, defenders must move just as fast—or faster—by automating security and vulnerability management. He discusses the urgent need to modernize legacy systems, evolve FedRAMP toward machine-to-machine compliance verification, and embrace AI-driven vulnerability scanning.Subscribe and listen to the Fed Gov Today Podcast anytime on Apple Podcasts, Spotify, or at FedGovToday.com.