OWASP GenAI Security Training Course

Day 1: Fundamentals and Core Threats

Module 1: Introduction to the OWASP GenAI Security Project (1 hour)

Learning Objectives:

• Comprehend the progression from the OWASP Top 10 to GenAI-specific security issues
• Investigate the OWASP GenAI Security Project ecosystem and available resources
• Distinguish key disparities between traditional application security and AI security

Key Topics:

• Mission and scope overview of the OWASP GenAI Security Project
• Introduction to the Threat Defense COMPASS framework
• Insights into the AI security landscape and regulatory mandates
• AI attack surfaces compared to conventional web application vulnerabilities

Practical Task: Configuring the OWASP Threat Defense COMPASS tool and conducting an initial threat evaluation

Module 2: OWASP Top 10 for LLMs - Part 1 (2.5 hours)

Learning Objectives:

• Master the first five pivotal LLM vulnerabilities
• Grasp attack vectors and exploitation tactics
• Apply effective mitigation strategies

Key Topics:

LLM01: Prompt Injection

• Direct and indirect prompt injection methods
• Hidden instruction attacks and cross-prompt contamination
• Real-world examples: jailbreaking chatbots and circumventing safety protocols
• Defensive measures: input sanitization, prompt filtering, and differential privacy

LLM02: Sensitive Information Disclosure

• Extraction of training data and leakage of system prompts
• Analyzing model behavior for sensitive information exposure
• Privacy impacts and regulatory compliance factors
• Mitigation: output filtering, access controls, and data anonymization

LLM03: Supply Chain Vulnerabilities

• Third-party model dependencies and plugin security
• Compromised training datasets and model poisoning
• Vendor risk evaluation for AI components
• Secure model deployment and verification protocols

Practical Task: A hands-on lab demonstrating prompt injection attacks on vulnerable LLM apps and establishing defensive measures

Module 3: OWASP Top 10 for LLMs - Part 2 (2 hours)

Key Topics:

LLM04: Data and Model Poisoning

• Techniques for manipulating training data
• Altering model behavior through poisoned inputs
• Backdoor attacks and data integrity checks
• Prevention: data validation pipelines and provenance tracking

LLM05: Improper Output Handling

• Insecure processing of content generated by LLMs
• Code injection via AI-generated outputs
• Cross-site scripting risks through AI responses
• Output validation and sanitization frameworks

Practical Task: Simulating data poisoning attacks and putting robust output validation mechanisms in place

Module 4: Advanced LLM Threats (1.5 hours)

Key Topics:

LLM06: Excessive Agency

• Risks of autonomous decision-making and boundary breaches
• Management of agent authority and permissions
• Unintended system interactions and privilege escalation
• Implementing guardrails and human oversight controls

LLM07: System Prompt Leakage

• Vulnerabilities involving system instruction exposure
• Disclosure of credentials and logic through prompts
• Techniques for extracting system prompts
• Protecting system instructions and external configurations

Practical Task: Designing secure agent architectures with proper access controls and monitoring

Day 2: Advanced Threats and Implementation

Module 5: Emerging AI Threats (2 hours)

Learning Objectives:

• Understand state-of-the-art AI security threats
• Implement advanced detection and prevention techniques
• Design resilient AI systems to withstand sophisticated attacks

Key Topics:

LLM08: Vector and Embedding Weaknesses

• Vulnerabilities in RAG systems and vector database security
• Embedding poisoning and similarity manipulation attacks
• Adversarial examples in semantic search
• Securing vector stores and implementing anomaly detection

LLM09: Misinformation and Model Reliability

• Detection and mitigation of hallucinations
• Bias amplification and fairness considerations
• Mechanisms for fact-checking and source verification
• Content validation and integration of human oversight

LLM10: Unbounded Consumption

• Resource exhaustion and denial-of-service attacks
• Strategies for rate limiting and resource management
• Cost optimization and budget controls
• Performance monitoring and alerting systems

Practical Task: Constructing a secure RAG pipeline with vector database protection and hallucination detection

Module 6: Agentic AI Security (2 hours)

Learning Objectives:

• Understand the specific security challenges of autonomous AI agents
• Apply the OWASP Agentic AI taxonomy to real-world systems
• Implement security controls for multi-agent environments

Key Topics:

• Introduction to Agentic AI and autonomous systems
• OWASP Agentic AI Threat Taxonomy: covering Agent Design, Memory, Planning, Tool Use, and Deployment
• Security and coordination risks in multi-agent systems
• Attacks involving tool misuse, memory poisoning, and goal hijacking
• Securing agent communication and decision-making processes

Practical Task: A threat modeling exercise applying the OWASP Agentic AI taxonomy to a multi-agent customer service system

Module 7: Implementing OWASP Threat Defense COMPASS (2 hours)

Learning Objectives:

• Master the practical use of Threat Defense COMPASS
• Integrate AI threat assessment into organizational security programs
• Formulate comprehensive AI risk management strategies

Key Topics:

• In-depth analysis of the Threat Defense COMPASS methodology
• Integration of the OODA Loop: Observe, Orient, Decide, Act
• Mapping threats to MITRE ATT&CK and ATLAS frameworks
• Creating AI Threat Resilience Strategy Dashboards
• Integration with existing security tools and processes

Practical Task: Performing a complete threat assessment using COMPASS for a Microsoft Copilot deployment scenario

Module 8: Practical Implementation and Best Practices (2.5 hours)

Learning Objectives:

• Design secure AI architectures from the ground up
• Implement monitoring and incident response for AI systems
• Establish governance frameworks for AI security

Key Topics:

Secure AI Development Lifecycle:

• Security-by-design principles for AI applications
• Code review practices for LLM integrations
• Testing methodologies and vulnerability scanning
• Deployment security and production hardening

Monitoring and Detection:

• AI-specific logging and monitoring requirements
• Anomaly detection for AI systems
• Incident response procedures for AI security events
• Forensic and investigation techniques

Governance and Compliance:

• AI risk management frameworks and policies
• Regulatory compliance considerations (GDPR, AI Act, etc.)
• Third-party risk assessment for AI vendors
• Security awareness training for AI development teams

Practical Task: Designing a comprehensive security architecture for an enterprise AI chatbot, including monitoring, governance, and incident response procedures

Module 9: Tools and Technologies (1 hour)

Learning Objectives:

• Evaluate and implement AI security tools
• Understand the current landscape of AI security solutions
• Build practical detection and prevention capabilities

Key Topics:

• The AI security tool ecosystem and vendor landscape
• Open-source security tools: Garak, PyRIT, and Giskard
• Commercial solutions for AI security and monitoring
• Integration patterns and deployment strategies
• Criteria for tool selection and evaluation frameworks

Practical Task: A hands-on demonstration of AI security testing tools and implementation planning

Module 10: Future Trends and Conclusion (1 hour)

Learning Objectives:

• Understand emerging threats and future security challenges
• Develop strategies for continuous learning and improvement
• Create action plans for organizational AI security programs

Key Topics:

• Emerging threats: deepfakes, advanced prompt injection, and model inversion
• Future developments and roadmap of the OWASP GenAI project
• Building AI security communities and knowledge sharing
• Continuous improvement and threat intelligence integration

Action Planning Exercise: Developing a 90-day action plan for implementing OWASP GenAI security practices in participants' organizations