On October 22, OpenAI and Google Cloud partnered to provide enterprise AI models, while NVIDIA announced a foundation for ...

What happens when your AI agents start outnumbering your employees 144 to 1? You've got a serious identity crisis on your ...

A METR employee and a Redwood Research employee spent three weeks stress-testing OpenAI's internal misalignment monitoring system, discovering a number of novel vulnerabilities.

China hackers agentic AI attack breached 2,250 Microsoft 365 tenants by exploiting a stolen ConnectWise Cloud Backup ...

The rulebook for what agents may touch is moving beneath the software layer, hours before AGNTCon opens with no shared spec ...

The models Astra and Sol misinterpreted instructions during internal use, which has serious implications for alignment ...

SentinelOne introduced Singularity Agentic Detection and Response, a platform designed to monitor and enforce AI agent behavior and detect and stop unintended or malicious activity in ...

OpenAI disclosed a pair of safety incidents that involved AI agents weakening their safety monitoring systems during internal ...

OpenAI said on Tuesday that two of its AI agents weakened parts of an internal monitoring system designed to detect misbehavior, the latest case of AI agents acting in ways their developers did not ...

We will check Azure VM extension provisioning failures by dividing them into Agent, handler, and guest processing.The target ...

New features in Microsoft 365 E7 and Project Perception will assess each agent's risk, with the ability to block it from ...

An AI agent may remember requesting a refund, but only the system of record can confirm it happened. Safe actions depend on ...