WA3556
Understanding Generative AI Risks: Security, Ethics, and Social Implications Training
As Generative AI (GenAI) becomes more widespread and impacts critical processes, understanding and managing its risks is crucial for successful operations. This risks of Generative AI training teaches attendees how to spot the dangers, risks, and vulnerabilities of GenAI.
Participants learn to methodically examine GenAI systems for potential hazards and identify mitigations for those risks. In addition, attendees gain industry-leading resources to stay updated in this rapidly evolving area.
Course Details
Duration
1 day
Prerequisites
This course is designed for personnel responsible for identifying, assessing, and managing the risks of Generative AI in their organization. It assumes they understand how Generative AI functions at a workflow level, including core steps in the training and prediction process.
Skills Gained
- Grasp the fundamental ethical principles of Responsible AI, such as fairness, accountability, and transparency, and how they affect real-world AI scenarios
- Discover how AI risks emerge from violations of Responsible AI principles
- Understand industry-standard categorizations and mitigations of AI risks
- Differentiate risks of Generative AI and the cybersecurity risks posed by it
- Identify potential vulnerabilities to AI systems and their defenses
Course Outline
- AI Ethics and Responsibility
- What is an AI system?
- The AI System Lifecycle
- Common AI Actors
- Principles of AI Ethics
- Safe
- Secure & Resilient
- Explainable & Interpretable
- Privacy-Enhanced
- Fair
- Accountable & Transparent
- Valid & Reliable
- Cybersecurity Triad & The Fallout of Failure
- GenAI Risks & Mitigations
- Nefarious Information (e.g., CBRN)
- Hallucinations
- Dangerous or Violent Recommendations
- Data Privacy
- Environmental Impacts
- Human-AI Configuration (e.g., workforce impact)
- Information Integrity (e.g., misinformation)
- Information Security
- Intellectual Property
- Obscene, Degrading, and Abusive Conduct
- Toxicity
- Bias
- Homogenization
- Supply Chain Integration
- GenAI Cybersecurity – Top 10 Vulnerabilities & Defenses
- OWASP LLM Top 10
- Prompt Injection
- Insecure Output Handling
- Training Data Poisoning
- Denial of Service
- Supply Chain Vulnerabilities
- Sensitive Information Disclosure
- Excessive Agency
- Overreliance
- Model Theft
- GenAI Cybersecurity – Tactics, Techniques, and Mitigations
- MITRE ATLAS
- Reconnaissance
- Resource Development
- Gaining Access
- Execution
- Persistence
- Privilege Escalation
- Defense Evasion
- Credential Access
- Discovery
- Collection
- ML Attack Staging
- Exfiltration
- Impact
- Conclusion
- Frontier Threats
- Additional Resources
- Responsibility Matters
