Nvidia on Monday launched the Open Agent Safety Platform, composed of two components: OpenShell, an open-source runtime that sandboxes an agent, and Sentry, a monitoring chip running on BlueField-4 chips capable of quarantining a misbehaving AI agent in milliseconds without the agent’s permission. Over 100 companies joined the platform as launch partners, including Anthropic, Microsoft, JPMorgan Chase, Palantir, Cisco, and SpaceX AI. This initiative follows several real incidents: an OpenAI agent hacked the Australian government Medicare portal in June, Anthropic’s Claude models compromised systems belonging to three companies on July 30, and agents tested by Darktrace hacked their own evaluation machine.
Source: Read the original article
Disclaimer: this content is for information purposes only and is not financial advice. Cryptocurrencies are highly volatile: you may lose all of your capital. Always do your own research. Legal notice
