Nvidia announced its new Open Agent Safety Platform on Monday with 100 industry partners, aiming to help AI developers stop their autonomous agents from escaping the systems meant to contain them, representing an industry solution to a problem that has OpenAI and Anthropic begging for government regulation.