Nvidia, Palantir Technologies, and Booz Allen Hamilton are limiting or ceasing the use of advanced AI models from Anthropic and OpenAI. The companies cite concerns that sensitive corporate data and intellectual property could be exposed through user prompts.
The firms are demanding stricter guarantees regarding data handling and privacy. Specific concerns include unclear data retention policies and the potential use of input metadata.
Palantir has reportedly secured zero-data-retention assurances from Anthropic to address these security fears.