Anthropic has partnered with Palantir and Amazon Web Services to incorporate its Claude AI models into U. S. intelligence and defense agencies. Claude, similar to ChatGPT’s language models, will use Palantir's platform and AWS hosting for data processing and analysis. However, the partnership has raised concerns about contradicting Anthropic’s "AI safety" principles. Critics, including former Google AI ethics co-head Timnit Gebru, have criticized the collaboration with Palantir on social media. Claude will operate within Palantir's Impact Level 6 environment, a secure system for handling national security data up to the "secret" classification.

This collaboration reflects a growing trend of AI companies pursuing defense contracts, similar to Meta and OpenAI. Claude's tasks will include analyzing complex data quickly, identifying patterns, and simplifying document reviews, while human officials maintain decision-making control. The partnership builds on Claude's integration into AWS GovCloud, with Anthropic expanding its operations in Europe and seeking funding at a $40 billion valuation, largely backed by Amazon. Despite its emphasis on ethical AI development, Anthropic's defense collaboration has drawn criticism, as it appears to conflict with its publicly promoted ethical stance. The partnership also ties Anthropic to Palantir, a controversial company involved in military AI projects like the Maven Smart System. While Anthropic’s terms of service limit and regulate government use of Claude, concerns persist about potential misuse and the AI models’ tendency to confabulate, which could be problematic with sensitive government data. These concerns highlight the risks associated with AI’s growing role in defense sectors, as noted by critics like Victor Tangermann.