MLOps.community The Evolution of AI in Cyber Security // Jeff Schwartzentruber // #344
36 snips
Nov 4, 2025 Jeff Schwartzentruber, a Senior Machine Learning Scientist at eSentire, dives into the evolving landscape of AI in cybersecurity. He reveals the shift from signature-based detection to dynamic anomaly detection, tackling issues like alert fatigue in Security Operations Centers. The conversation explores the risks associated with AI agents, including prompt injections and the need for visibility. Jeff highlights how defenders and attackers use Generative AI, emphasizing the importance of maintaining organizational truth amid rising deception risks.
AI Snips
Chapters
Transcript
Episode notes
From Signatures To Dynamic Detection
- Signature-based detection works but breaks when malware mutates its hash or exact indicators.
- Thresholding and dynamic anomaly detection began filling the gap by profiling users and activity patterns.
Anomaly Detection's Scaling Problem
- Profiling and anomaly detection scale detection by baselining individual behavior instead of static rules.
- High cardinality and false positives remain major operational challenges for anomaly systems.
GenAI Turns Logs Into An NLP Problem
- GenAI complicates detection because telemetry is unstructured and normalization varies across vendors.
- Security becomes an NLP-like engineering problem when parsing diverse logs and extracting context.