Topic · 2 posts
AI security
Prompt injection, model risk, and securing LLM systems.
All topics- Prompt Injection: The Trust-Boundary Bug in LLM Applications How direct and indirect prompt injection cross trust boundaries, and how to limit what a manipulated model can expose or do. 6 min read
- AI in IT Infrastructure and Cybersecurity: Where It Helps and Where It Fails A practical framework for evaluating AI in monitoring, threat detection, investigation, and response. 7 min read