Gecko机器人与NVIDIA应添加AI代理安全和控制

全文内容 Full Text

Gecko is using NVIDIA’s new agentic AI security layer in its inspection robots. Source: Gecko Robotics NVIDIA Corp. today announced its Open Agent Safety Platform, which consists of the OpenShell open-source software and Sentry reference system design. They are intended to strengthen governance and control across hardware, compute, and software for robots that run AI agents. Among the more than 100 organizations working with NVIDIA’s technology is Gecko Robotics Inc. After recent incidents including AI agents from OpenAI going around application-level controls to attack Hugging Face , AI and robotics developers have been working on traceability, accountability, and preventative measures. “As we continue to discover the frontier of AI capabilities, we must accelerate discovery at the frontier of AI safety,” stated Jensen Huang, CEO of NVIDIA. “Safety and security require full-stack engineering. NVIDIA Open Agent Safety Platform brings together industry, researchers, and public-sector organizations to share best practices, align on resolution methods, and foster international cooperation.” Editor’s note: Physical AI is among the session topic tracks at RoboBusiness 2026, which will be on Oct. 20 and 21 in Santa Clara, Calif. Register now to attend. Register now and help us celebrate 20 years of RoboBusiness! NVIDIA OpenShell and Sentry enforce policy OpenShell provides a secure runtime boundary that traces all actions and enforces policy as AI agents run on NVIDIA Vera CPUs. It provides deterministic governance for agent execution, access, and where inference goes, according to NVIDIA. As open-source software, OpenShell can be extended to work with third-party compute platforms, including those from Arm and Intel, added the Santa Clara, Calif.-based company . NVIDIA Sentry adds an out-of-band watchdog that runs on BlueField-4 data-processing units (DPUs) to continuously monitor agent behavior. It monitors outcomes against policies, the company said . Sentry can quarantine agents that attempt to move outside their boundaries in milliseconds. “For the agent economy to grow, we need a trusted foundation across silicon and software,” said Justin Boitano, vice president of enterprise AI at NVIDIA, during a press briefing. “We want to engage everybody to advance a new layer of agent security.” “The independent trust domain is like how self-driving cars work,” he explained. “The primary system runs perception, and a safety island ensures that the overall system fails safely. The same principles apply to frontier AI systems.” OpenShell can offload security onto silicon or DPUs as needed, and Sentry watches reasoning traces. A Policy Proover in NVIDIA Open Agent Safety Platform is deterministic and runs at a higher speed. “If you write a policy that says an agent can’t read code from Github, an agent could spawn separate subagents to read it and then post that information, with fleets superseding global policies,” said Boitano. “The Policy Proover validates the master decision tree, looking at combined file and network access to find unintended exfiltration. It comes out of AWS’s S3 research to create verifiable policies.” “We see long-term agent behavior to be a different problem than sandboxing,” he added. “OpenShell is projecting the intent of a policy into infrastructure.” The NVIDIA Open Agent Safety Platform includes OpenShell and Sentry. Source: NVIDIA While recent security incidents have occurred entirely within software, Gecko Robotics said they revealed a potential problem with all AI agents, including robots. The company has used NVIDIA OpenShell to explore how enforceable boundaries can keep AI-powered robots operating within human-defined permissions. Gecko said its goal is to establish the safeguards needed to deploy greater autonomy responsibly. “As Jensen says, safety is an engineering problem, not a legal one,” said Jake Loosararian, co-founder and CEO of Gecko Robotics. “The idea that losing control of AI