A focused series of one-hour technical sessions covering the practical security challenges associated with artificial intelligence, machine learning systems, generative AI applications, and large language models. The sessions help technical professionals understand how AI systems are attacked, how their risks differ from traditional application security, and how security controls can be integrated across development, deployment, testing, and incident response activities.
The sessions address the security of AI applications across the full technical lifecycle—from training data, models, dependencies, and development pipelines to APIs, agent permissions, production monitoring, and incident response. Each session focuses on a specific technical risk or security practice that participants can relate directly to real AI-enabled systems and enterprise use cases.
Eight focused technical sessions covering the AI security lifecycle—from application architecture and prompt attacks to secure development, assurance, and incident response.
Security Architecture, Attack Surface & Defensive Controls for Generative AI Applications
This session introduces the security architecture of generative AI and LLM applications, covering risks across prompts, models, APIs, data sources, plugins,…
Direct, Indirect & Agent-Based Prompt Injection Risks
This session examines how attackers manipulate AI applications through direct and indirect prompt injection, bypass system instructions, expose sensitive data,…
Protecting Knowledge Sources, Vector Databases & Retrieval Pipelines
This session explores the security risks of Retrieval-Augmented Generation systems, including poisoned documents, unauthorized data retrieval, weak access…
STRIDE, Trust Boundaries, OWASP LLM & MITRE ATLAS
This session shows how to identify and prioritize threats across AI applications using data-flow diagrams, trust boundaries, STRIDE, OWASP guidance, MITRE…
Securing Models, Datasets, Dependencies & Third-Party AI Components
This session examines risks introduced by external models, datasets, open-source packages, model repositories, compromised dependencies, backdoored components,…
Embedding Security Across AI Development and Deployment Pipelines
This session covers how to integrate security controls into AI development, MLOps, and CI/CD pipelines through dependency analysis, model validation, artifact…
Adversarial Testing, LLM Assessments & AI Security Validation
This session introduces practical methods for testing AI applications, including prompt attacks, API assessments, data-leakage testing, tool-abuse scenarios,…
Detection, Containment, Investigation, Recovery & Model Rollback
This session examines how organizations should respond to AI-specific incidents such as model poisoning, data exfiltration, compromised AI services, malicious…