Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

AI agents don't think. They pattern-match. Critical to understand: Generative AI (ChatGPT, Claude, etc.) does NOT reason like humans. It: The API Security problem: When you give an AI agent access to an API, it will: AI agents can't reason. They recreate patterns based on weights. You need to be very careful: data in, data out. Practical example: text User: "Show me the account balance for user" AI agent → calls GET /api/account/123 API → returns { balance: 5000, name: "John", SSN: "123-45-6789" } AI agent → outputs EVERYTHING to user (including SSN!)

Software release cycles are now too fast for traditional security tools. Rapid iterations and reliance on open-source and cloud-native tech increase vulnerabilities, challenging AppSec teams to keep up. Attackers are taking advantage, targeting applications and exploiting misconfigurations, excessive permissions, and vulnerable plug-ins.

How to Enable GLM 4.7 Beta in Windsurf IDE.

In Episode 7 of Corelight DefeNDRs, join me, Richard Bejtlich, as I sit down with Dr. Keith Jones, Corelight's principal security researcher, to discuss the practical applications of AI in enhancing network security. We delve into how large language models (LLMs) can assist in cleaning up documentation and generating Zeek scripts, sharing insights from our extensive experience in incident response and coding. Keith reveals the challenges and successes he has encountered using LLMs to streamline processes, including their role in analyzing MITRE techniques.