Veza launches AI Agent Security amidst growing risks

As reported by Silicon Angle, Veza Inc. has announced the launch of AI Agent Security, a new product designed to help organizations manage and secure artificial intelligence agents at an enterprise level. This offering aims to provide critical visibility and control over AI agents as businesses increasingly adopt AI technologies.

AI Agent Security, built on Veza's Access Graph, offers unified visibility across various AI platforms, enabling security teams to identify existing AI agents and understand human control over them. The product addresses emerging security risks like prompt injection attacks, where malicious actors can manipulate AI agents to access confidential data or perform unauthorized actions. It achieves this by discovering and classifying agent identities, visualizing their connections to data resources and large language models, and enforcing the principle of least privilege. The service also maps AI agent access to human identities, allowing governance over deployment and interaction, and supports compliance with security recommendations for access audits.

The introduction of AI Agent Security comes at a time when the rapid deployment of AI agents outpaces organizations' ability to govern them, creating a "Wild West" scenario. Veza's solution aims to bring order by unifying discovery, governance and control, emphasizing the critical role of identity in understanding AI access. This launch is particularly notable as Veza is in the process of being acquired by ServiceNow.

Source: Silicon Angle