"Exabeam is launching five new features that collectively cover the entire agentic attack surface. AI behavior baselining builds dynamic behavior profiles for users and their AI agents, tracking patterns in request volumes, token usage, tool calls, and outbound traffic."
"Deviations in AI agent behavior are immediately flagged to security teams, ensuring that human intervention occurs as quickly as possible. This proactive approach is essential in addressing the risks posed by AI agents in business processes."
"The introduction of prompt and model abuse detection is designed to identify prompt injection, model manipulation, and the exploitation of tools. This detection library is five times larger than the previous version, enhancing the capability to monitor AI agent behavior."
Exabeam is enhancing Agent Behavior Analytics with support for ChatGPT, Microsoft Copilot, and Google Gemini. Five new features will monitor AI agent behavior for anomalies, detect prompt injection, and verify access rights. These features address significant AI risks and improve integration with New-Scale and LogRhythm SIEM platforms. The expansion aims to mitigate security gaps caused by invisible AI agents, which can misuse privileges and lead to data breaches. The new features include behavior baselining, prompt and model abuse detection, and identity monitoring.
Read at Techzine Global
Unable to calculate read time
Collection
[
|
...
]