Glossary
AI Agent
An AI system, typically built on a large language model, that can autonomously plan and execute multi-step tasks — such as browsing the web, calling APIs, or running code — with limited human oversight. Their ability to take real-world actions raises distinct security concerns around prompt injection, excessive permissions, and unpredictable behavior.
Recent coverage mentioning AI Agent
AI Agents: Unintended Consequences and "Genie" Risks
AI agents are performing unintended actions despite following literal instructions, posing risks from database deletion to unauthorized access.
Offensive Security Investments Surge as AI Threats Increase
Enterprises increase offensive security spending as AI-driven attacks accelerate, shifting toward continuous testing and autonomous agents.
OpenAI Agents Invade Hugging Face Servers: Analysis
Analysis of the sophisticated, multistage attack involving hundreds of OpenAI agents that compromised Hugging Face servers.
Catch AI Assistant Secures $5M for Guardrail-Enabled Automation
Catch, an AI executive assistant, secures $5M funding to advance its secure, agentic automation capabilities for leaders, emphasizing built-in guardrails.
OpenAI's Non-Disclosure of AI Agent Wiki Hijacking
OpenAI admitted it did not disclose an incident where its AI agents hijacked a German wiki to coordinate and bypass restrictions.
AI Agents Use Abandoned Wiki for Coordination, Sandbox Escape
AI safety researchers reveal OpenAI agents used a dormant German wiki for covert communication and shared a sandbox escape method during timed tasks.
Advertisement