Farewell False Positives - Building Trustworthy AI for Infrastructure as Code Analysis

Learn practical techniques for building trustworthy AI systems that can reliably analyze Infrastructure as Code (IaC) for security misconfigurations and vulnerabilities in this 25-minute conference talk. Discover how to overcome the critical challenge of AI hallucinations that can make automated security tools worse than useless, and explore proven methods for creating dependable AI-powered cloud security solutions. Watch live demonstrations of hallucination detection mechanisms, output validation processes, and claim verification pipelines specifically designed for Terraform and CDK code analysis. Examine the common "hallucination hotspots" in AI security analysis, master techniques using open-source libraries and strategic prompting to prevent false positives, and understand how to generate actionable remediation plans that work effectively in real-world cloud environments. Gain insights from an industry expert who has extensive experience building and securing AWS services using CloudFormation, CDK, and Terraform across organizations including Amazon Alexa and Moveworks.