Nvidia has introduced a new security platform aimed at controlling artificial intelligence agents to prevent unauthorized actions, addressing rising concerns about AI systems overstepping their intended boundaries. The Open Agent Safety Platform, as it is called, features open-source software named OpenShell, which enables developers to set and verify limits on AI agents’ permissions.
The platform also includes a security component known as Sentry, which continuously monitors AI agent activity. According to Nvidia, Sentry can identify suspicious behavior and quickly isolate any agent attempting to operate beyond its designated limits. This technology is particularly relevant in light of recent incidents where AI systems have autonomously accessed or interacted with external organizations without authorization.
More than 100 organizations, including major players like Microsoft, Perplexity, Accenture, and JPMorgan Chase, are already utilizing the platform. Notably, OpenShell’s open-source design allows it to be extended to work with computing platforms from other chip manufacturers, such as Arm and Intel.
In addition to unveiling the new security platform, Nvidia’s board has approved a significant expansion of its stock repurchase program, authorizing an additional $150 billion for buybacks, bringing the total to $235 billion. This move represents one of the largest corporate stock buyback programs in U.S. history. Nvidia CEO Jensen Huang stated that the company’s strong cash generation enables it to both invest in technology advancements and return capital to shareholders.