Stephen Sims, Curriculum Lead & Faculty Fellow, SANS Institute The most productive way to think about AI in cybersecurity is not as a replacement for people, but as an amplifier of human capability. It can accelerate analysis, expand testing coverage, summarize complex information, and help practitioners explore more possibilities. Yet speed without judgment can create noise, false confidence, and dangerous decisions. This talk explores the balance between automation and accountability. It examines why AI systems frequently overstate severity, why incomplete context leads to flawed conclusions, and why development teams may become overwhelmed by findings that have not been properly triaged or validated. It also considers the risks of delegating authority to local agents that can access files, credentials, browsers, code, and internal systems. The goal is to show how organizations can move quickly with AI without removing the human responsibility that makes security trustworthy. Register for the 2027 Secure Connections at: https://secure-connections.att.com/ Subscribe to our channel! http://youtube.com/att At AT&T, we believe that Connecting Changes Everything. And we love connecting you to the people, passions and experiences you love most. That means bringing you exclusive content with your favorite athletes, artists and celebrities, tutorials, product demos, new device features, and more stories from the heart of AT&T. Visit the AT&T Website: https://www.att.com Like AT&T on Facebook: http://facebook.com/att Follow AT&T on X: http://x.com/att Follow AT&T on Instagram: http://instagram.com/att Follow AT&T on LinkedIn: https://linkedin.com/company/att/ For Customer Care issues: https://x.com/ATTHelp For Business Customer Care issues: https://x.com/ATTBusinessCare

 221      3