AI Security Podcast How Microsoft Uses AI for Threat Intelligence & Malware Analysis
40 snips
Oct 18, 2025 Thomas Roccia, a Senior Threat Researcher at Microsoft specializing in AI applications for malware analysis, discusses groundbreaking concepts like the 'Indicator of Prompt Compromise' (IOPC). He shares insights on his open-source projects, including NOVA, a tool to detect malicious prompts. The conversation explores using AI to track complex crypto laundering schemes, simplifying reverse engineering, and how AI enhances threat intelligence. Roccia also highlights the shift in skill accessibility, where advanced tasks become manageable for more professionals.
AI Snips
Chapters
Transcript
Episode notes
Layered Prompt Detection Strategy
- Use layered detection for adversarial prompts: keywords, regex, semantic embeddings, and optionally an LLM judge.
- Combine conditions (AND/OR) to increase detection confidence and reduce reliance on a single method.
Harden RAG With Hybrid Retrieval
- Improve RAG retrieval with hybrid search (keyword + semantic) and careful chunking when ingesting data.
- Tune chunk size and parsing to preserve cross-document context and reduce missed references.
Scan MCP Servers Before Use
- Scan Model Context Protocol (MCP) servers before use to identify weaponized tool descriptions and risky prompts.
- Apply rules and score risk proactively to avoid consuming malicious MCP resources.