Exabeam Boosts AI Security: New Tools for Autonomous Agents

Cybersecurity company Exabeam has introduced new features to help security teams manage risks from autonomous AI agents. This update expands several Exabeam tools, including Agent Behavior Analytics and Threat Center. The company has doubled its AI-focused behavioral detections to 90. These detections can identify suspicious prompt activity, unauthorized autonomous agents, and abnormal AI usage patterns. They also help detect "Denial of Wallet" attacks that generate excessive costs through AI services. Exabeam's platform now supports more leading AI platforms, adding Anthropic Claude to its coverage of OpenAI ChatGPT, Google Gemini, Microsoft Copilot, and GitHub Copilot. This helps organizations see which AI tools are in use. Exabeam has also launched Observra, an open-source library to standardize activity data from AI agents. This helps feed security data into existing security platforms. The bottom line: These advancements aim to help organizations secure their operations as they increasingly use autonomous AI agents.