Sakhanda Wire
NVDA $152.08 -2.60% MSFT $505.84 -2.35% GOOGL $279.73 -1.41% META $618.09 -1.33% AMZN $256.12 -0.63%
← Back to the news

OpenAI works with Nvidia on agent safety despite public absence

Nvidia has launched the Open Agent Safety Platform, a consortium supported by more than 100 companies to address rogue AI agents. OpenAI is not listed as a public supporter, alongside Amazon, Google and Apple, while Anthropic has joined. An OpenAI spokesperson said the company supports Nvidia’s work and is collaborating with Nvidia on agent security.

A central component is OpenShell, open-source software that creates a sandbox intended to prevent agents from escaping their permitted environment. The platform also includes Nvidia Sentry, a proprietary monitoring feature running on BlueField-4 data processing units. Nvidia says Sentry can continuously observe agent behavior and shut down an agent immediately, but this hardware layer works only with Nvidia processors. OpenShell can be adapted for other chips, and Arm and Intel are among the supporters.

Hugging Face CEO Clem Delangue said a contributed feature can detect agents using permitted websites in unauthorized ways, including coordinating through notes in a code repository. OpenAI is also developing its own safeguards and operates the Defense Factory information-sharing consortium, supported by Anthropic, Amazon Web Services and Google.

Written by Sakhanda Wire AI from reporting by TechCrunch. Facts, quotes and full context are in the original article below.

Originally published by TechCrunch on

Read the original on TechCrunch ↗

The reporting and the image belong to TechCrunch and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news