Anthropic Announces 'Claude Gov' AI Model for Defense and Intelligence Agencies

Anthropic has announced 'Claude Gov,' an AI model designed specifically for the US defense and intelligence agencies. The model is loosely regulated for government agencies, trained to improve the accuracy of analyzing classified information, and is 'already deployed at the highest levels of national security.'

Claude Gov Models for US National Security Customers\Anthropic
https://www.anthropic.com/news/claude-gov-models-for-us-national-security-customers

Anthropic launches Claude Gov for military and intelligence use | The Verge
https://www.theverge.com/ai-artificial-intelligence/680465/anthropic-claude-gov-us-government-military-ai-model-launch

According to Anthropic, Claude Gov is designed specifically for government needs such as threat assessment and intelligence analysis. Anthropic says that it has undergone rigorous safety testing, as have all Claude models, but it also has specific uses for national security operations, and it appears to have less restrictions on classified information than the general Claude.

Claude Gov also provides a deeper understanding of documents and information in intelligence and defense contexts, addresses words and phrases essential to national security operations, and improves the understanding and interpretation of complex cybersecurity data for intelligence analysis.

Anthropic's usage policy states that users may not 'create or facilitate the exchange of illegal or highly regulated weapons or goods,' which includes using Anthropic's products or services to 'manufacture, modify, design, sell, or distribute weapons, explosives, hazardous materials, or other systems intended to harm or take life.'

However, in 2024, Anthropic added an exception to its usage policy stating, 'Anthropic may enter into contracts with certain government customers to customize usage restrictions to align with those customers' public missions and legal authorities if Anthropic determines that the contractual usage restrictions and applicable safeguards are sufficient to mitigate the potential harms addressed in this Usage Policy.' This case appears to fall under this exception.

This is not the first time that an AI company has collaborated with a government agency; OpenAI announced 'ChatGPT Gov' in January 2025.

OpenAI announces government-only version of ChatGPT 'ChatGPT Gov' - GIGAZINE

In addition, Scale AI, an AI startup that provides training data to OpenAI, Google, Meta and others, has signed a contract for an AI agent program aimed at developing military plans for the US military.

Anthropic explained that Claude Gov 'builds on our commitment to providing responsible and secure AI solutions to U.S. national security customers, leveraging custom models built to address the unique needs of classified environments.'