News for today:
- [Absolute Security] Firmware-embedded security with NVIDIA. Absolute Security announced an agreement with NVIDIA to embed security, endpoint management, and resilience capabilities in NVIDIA firmware, to add a layer of protection to AI workstations. This integration will help enterprise security and IT leaders ensure their AI developers’ autonomous agents function within the boundaries of security and compliance policies, and that their devices remain defended against threats, protected from risk, and able to quickly resume operations following cyber disruptions. This innovation will be available to any OEM partner building the NVIDIA DGX Station AI Supercomputer on Linux and Windows. Absolute Security
- [CYE] Multi-agent AI platform. CYE launched a new agentic platform for the continuous analysis of exposure, mitigation options, and remediation pathways; the platform utilizes and leverages multiple agents. As the environment evolves, Cye tracks changes, identifies new gaps, and helps security teams continuously improve their security posture. Grounded in each organization’s security data, environment and context, they accelerate the path from insight to action, helping teams make faster, more informed decisions and continuously reduce exposure. CYE
- [Defakto] OpenID membership. Defakto joined the OpenID Foundation to support the development of future open identity standards. Defakto’s participation in the OpenID Foundation builds on years of leadership in open identity standards. Built on SPIFFE, OAuth and OpenID Connect, and emerging standards such as WIMSE. Defakto’s platform extends standards-based identity across infrastructure, applications, automation and AI without introducing another proprietary identity layer. Defakto
- [BeyondTrust] AI agents acting beyond intent. BeyondTrust GA’d AI Agent Security on its identity security platform, offering security, visibility, and control over the actions of AI agents. The release enables organizations to distinguish AI agent activity from human activity and [enforce] what agents are permitted to do at runtime, before unauthorized actions occur, without restricting the permissions employees need to perform their jobs. Research by BeyondTrust shows multiple ways that AI agents can and are going beyond what they are supposed to be doing. BeyondTrust
- [IRONSCALES] AI agent security incidents. New research from IRONSCALES on the same topic (as BeyondTrust) highlights the extent of agentic AI-related security incidents. The data: 77% of organizations say an AI agent or tool acted on instructions or content that turned out to be malicious, misleading, or unauthorized in the past 12 months. More than half (57%) report a confirmed case. IRONSCALES
- [SentinelOne] Expanded partnership with N-able. SentinelOne and N-able expanded their cybersecurity partnership beyond endpoints to cover additional attack areas. N-able and SentinelOne help organizations strengthen security across an ever-growing attack surface [including AI, cloud, identities, and managed response] by harnessing AI to deliver a comprehensive approach to cyber defense and overall business resilience. These capabilities span three key AI roles: prevention and containment, investigation and triage, and AI governance. SentinelOne