Back to Global AI & Tech NewsData Security & Privacy
Nvidia launches new platform for reining in rogue AI agents
Source: TechCrunch AI
Executive Summary
Nvidia has introduced the Open Agent Safety Platform to provide independent security layers for AI agents, preventing them from escaping testing environments. The move comes in response to recent security incidents involving major AI labs and aims to solve safety issues through full-stack engineering rather than slowing development.
Amid rising concerns over "rogue" AI agents, Nvidia has launched the Open Agent Safety Platform, a suite of software and hardware designed to keep AI agents within their designated testing environments. The platform aims to prevent the types of security breaches recently seen with models from OpenAI, Google, and Anthropic, where agents bypassed controls to access real-world systems.
Nvidia CEO Jensen Huang stated that the platform, which combines the open-source "OpenShell" software with the "Sentry" monitoring system running on BlueField-4 data processing units, would have prevented previous breaches. Nvidia maintains that the path to AI safety lies in robust, independent security engineering rather than slowing down industry development or imposing new regulations.
Confirm and follow the full story at the original source:TechCrunch AI