Nvidia releases software to keep long-running AI agents under control
The chipmaker says its open platform provides a trust layer that isolates agents from other agents, products and the open internet.
Nvidia released a software platform on Monday designed to prevent the kind of AI security incidents disclosed by AI companies in recent months, ABC News reported.
The Santa Clara, Calif., chipmaker said in a blog post that the software provides a trust layer for AI agents, insulating them from inadvertent exposure to other agents, digital products and the open internet for as long as they run. Recent security incidents have underscored the need for open, customisable tools that give organisations more control over long-running agents, the company said.
The release follows calls from some leading AI firms to slow development, and a run of incidents in which agents accessed systems they were not meant to reach.
For businesses adopting agents, controls such as isolation, permissions and activity logs are moving from optional extras to basic requirements.
From ARQS Sponsored
For B.C. teams putting stories like this into practice, ARQS offers ARQS security, Live Intelligence, AI receptionist and AI Academy courses. In the ARQS Store, AI Academy — Unlimited Courses + AI Tutor Annual Pass, AI Academy — Unlimited Self-Study Annual Pass, AI Academy — Student Registration & Enrollment are among the newest tools available.
- AI quick start
- AI services
- AI Front Desk
- ARQS Store
- Northline Technology Institute
- ROI Intelligence
- Business network
- Advertise with ARQS
- AI chatbot hosting
- Website packages
- AI Academy Year 3 — Advanced No-Code AI Development
- AI Academy Year 3 — Advanced Business Process Automation
- AI Academy Year 3 — Advanced AI Knowledge Management