Securing AI and Bedrock guardrails

How do you move faster with AI and cloud innovation without losing control of security along the way? Recorded live from the show floor at AWS re:Invent in Las Vegas, this episode of Tech Talks Daily features a timely conversation with Kimberly Dickson, Worldwide Go-To-Market Lead for AWS Detection and Response Services. As organizations race to adopt agentic AI, modernize applications, and manage sprawling cloud environments, Kimberly offers a grounded look at why security must still sit at the center of every decision. Kimberly explains how her role bridges two worlds at AWS. On one side are customers dealing with prioritization fatigue, fragmented security signals, and growing pressure to do more with fewer resources. On the other hand, there are the internal service teams building products like Amazon GuardDuty, Amazon Inspector, and AWS Security Hub. Her job is to connect those realities, shaping services based on what customers actually struggle with day to day. That perspective sets the tone for a conversation focused less on hype and more on practical outcomes. We unpack how AWS thinks about security culture at scale, from infrastructure and encryption through to threat intelligence gathered across Amazon's global footprint. Kimberly shares how AWS uses large-scale honeypots to observe attacker behavior in real time, feeding that intelligence back into detection services while also working with governments and industry partners to take down active threats. It is a reminder that cloud security is no longer just about protecting individual workloads, but about contributing to a safer internet overall. The conversation also dives into new announcements from re:Invent, including the launch of AWS Security Hub, extended threat detection for EC2 and EKS, and the emergence of security-focused AI agents. Kimberly explains how these tools shift security teams away from manual investigation and toward faster, higher-confidence decisions by correlating risks across vulnerabilities, identity, network exposure, and sensitive data. The goal is clear visibility, clearer priorities, and remediation that fits naturally into existing workflows. We also explore how AWS approaches security in multi-cloud and hybrid environments, why foundational design principles still matter in an AI-driven world, and how open standards are helping normalize security data across vendors. Kimberly's reflections on re:Invent itself bring a human close to the episode, highlighting the pride and responsibility felt by teams building systems that millions of organizations depend on. As AI adoption accelerates and security teams are asked to keep pace without slowing innovation, what would it take for your organization to move faster while still trusting the foundations you are building on?