Anthropic unveils custom AI models for U.S. national security customers

Anthropic has announced the introduction of a new suite of AI models specifically designed for U.S. national security applications. These models, known as the ‘Claude Gov’ series, were developed in response to direct input from government clients to meet their operational challenges. Unlike Anthropic's models aimed at consumers and businesses, the Claude Gov series is tailored for tasks such as strategic planning, operational support, and intelligence analysis. According to a blog post from Anthropic, these models are currently being utilized by high-level national security agencies, with access restricted to personnel operating within classified settings. The Claude Gov models have undergone the same stringent safety evaluations as the company's other offerings. As Anthropic seeks to diversify its revenue streams, its engagement with government clients has intensified. In November, the company partnered with Palantir and AWS, Amazon's cloud computing branch, to market its AI solutions to defense sectors. The new Claude Gov models are designed to manage classified information more effectively, exhibit lower refusal rates when processing sensitive data, and demonstrate enhanced comprehension of documents relevant to intelligence and defense. Anthropic claims these models also possess improved capabilities in understanding critical languages and dialects for national security operations, as well as a refined ability to analyze complex cybersecurity data. Anthropic is not alone in vying for defense contracts; OpenAI is looking to strengthen its ties with the U.S. Defense Department, while Meta has announced plans to provide its Llama models to defense partners. Additionally, Google is developing a version of its Gemini AI for classified settings, and Cohere is collaborating with Palantir on deploying its AI solutions.