On September 14th, it was reported that Nvidia, Palantir, and Booz Allen Hamilton have requested OpenAI and Anthropic to provide new security guarantees and reduce or discontinue the use of their most advanced AI models. Among them, Palantir has prompted Anthropic to commit to providing an 'irrevocable' zero data retention (ZDR) guarantee for all models. Currently, some companies are shifting towards using self-developed AI models to handle sensitive operations.
