INGA CHERNY

BLOGS

Why AI Vulnerability Scanners Score 68% in the Lab and Fail in Production
LATEST2026-01-28

Why AI Vulnerability Scanners Score 68% in the Lab and Fail in Production

The gap between lab benchmarks and production reality reveals a fundamental flaw in how we evaluate AI security tools. Same model. Same task. Wildly different results.

#AI_SECURITY#BENCHMARKS#VULNERABILITY
Your Reasoning Model Is Thinking Too Hard. That's the Attack.
2025-12-25

Your Reasoning Model Is Thinking Too Hard. That's the Attack.

They crafted adversarial suffixes that force reasoning models to think 3 to 9 times longer than necessary.

#AI_SECURITY#RESEARCH
A Zero-Click Attack That Doesn't Need Your AI Assistant
2025-12-01

A Zero-Click Attack That Doesn't Need Your AI Assistant

A legitimate delete my data request. The kind companies are legally required to process.

#AI_SECURITY#RESEARCH
When a Single Word Can Hijack an AI
2025-11-27

When a Single Word Can Hijack an AI

A model can behave perfectly normally in almost every situation until it sees a specific phrase that activates something it learned during its earliest training.

#AI_SECURITY#RESEARCH
Your AI Code Improvements Are Making Things Worse
2025-11-24

Your AI Code Improvements Are Making Things Worse

A new peer-reviewed study analyzed 400 code samples through 40 rounds of AI improvements using four distinct prompting strategies.

#AI_SECURITY#RESEARCH#VIBE_CODING