Anthropic releases "Claude Gov" AI models for U.S. national security tasks, enhancing intelligence operations and partnering with Palantir and AWS for defense contracts.

On Thursday, Anthropic introduced specialized AI models specifically engineered for U.S. national security clients. The company unveiled "Claude Gov" models, developed in response to direct input from government customers to manage tasks such as strategic planning, intelligence analysis, and operational support. These custom models are reportedly already in use by U.S. national security agencies, with access limited to those involved in classified environments.

The Claude Gov models differ from Anthropic's consumer and enterprise versions, also known as Claude, in several key areas. They are reportedly adept at managing classified materials, are more accommodating when interacting with classified information, and are tailored to work with intelligence and defense documents. Furthermore, the models boast what Anthropic describes as "enhanced proficiency" in languages and dialects vital to national security operations.

Anthropic assures that the new models have undergone the same rigorous "safety testing" as all Claude models. In its quest for stable revenue streams, the company has been actively pursuing government contracts, partnering with Palantir and Amazon Web Services in November to market AI tools to defense clients.

Anthropic is not the first to offer specialized chatbot services for intelligence agencies. In 2024, Microsoft launched an isolated version of OpenAI's GPT-4 for the U.S. intelligence community following 18 months of development. This system, operating on a dedicated government-only network without Internet access, became operational for approximately 10,000 people within the intelligence community for testing and information retrieval purposes.