Anthropic launches Claude Gov, an AI model tailored for intelligence and defense agencies

Anthropic has now introduced Claude Gov, a new suite of large language models designed specifically for use by U.S. government defence and intelligence agencies. The AI models are built to handle classified data with fewer restrictions and offer more advanced context understanding in high-security environments.

According to the company, Claude Gov is “already deployed by agencies at the highest level of U.S. national security,” although it declined to disclose when the rollout began or which departments are using the technology. Access to the Claude Gov models is restricted to government agencies working with classified material.

The models are positioned as highly specialised tools for intelligence analysis and threat assessment. Unlike consumer-facing Claude models, which are trained to reject engagement with sensitive or confidential information, Claude Gov is designed to be more permissive when working with classified inputs. “They refuse less when engaging with classified information,” Anthropic stated in a blog post.

The company also claims these models offer superior comprehension of defence-related documents, proficiency in mission-critical dialects, and tailored capabilities for national security contexts. However, Anthropic emphasised that Claude Gov underwent the same rigorous safety testing as its public counterparts.

The launch of Claude Gov marks Anthropic’s strategic push into the competitive arena of government AI solutions, going head-to-head with OpenAI’s ChatGPT Gov, which debuted in January. OpenAI reported that more than 90,000 U.S. government employees had used its technology within the past year for everything from policy drafting to code generation.

Anthropic declined to share comparative usage figures, but confirmed its involvement with Palantir’s FedStart programme, an initiative supporting software vendors aiming to serve U.S. federal clients.

The release also reignites ongoing debates around the use of AI in government. Critics have raised concerns about the misuse of artificial intelligence in policing, surveillance, and social services. Facial recognition tools, predictive policing models, and algorithmic bias in welfare assessments have all come under fire for disproportionately impacting marginalised communities.

In response to such concerns, Anthropic reiterated its commitment to ethical guardrails, stating that its usage policy prohibits AI deployment for disinformation campaigns, weapons development, censorship systems, and malicious cyber operations. However, the company said it has introduced “contractual exceptions” tailored for specific government missions, aiming to “balance enabling beneficial uses with mitigating potential harms.”

Claude Gov joins a broader wave of AI deployments in government infrastructure. In March, training data provider Scale AI signed a deal with the U.S. Department of Defense for AI-powered military planning. Since then, Scale has also secured a five-year agreement with Qatar to digitise civil services.