Anthropic is expanding its artificial intelligence capabilities to the public sector, aiming to support government functions through advanced AI systems. Leveraging Amazon Web Services (AWS), the company’s AI models, Claude 3 Haiku and Claude 3 Sonnet, are now available to the US Intelligence Community (IC) and in AWS GovCloud.
This initiative allows government agencies to utilize Claude for various applications, including enhancing citizen services, streamlining document processes, and supporting better policymaking through data-driven insights. Future applications could involve disaster response coordination, public health initiatives, and optimizing energy grids. When used responsibly, AI has the potential to significantly transform government operations, contributing to peace and security.
Anthropic has tailored its offerings to meet the stringent security standards and unique requirements of government users. The company has adjusted its service agreements, providing a set of contractual exceptions for beneficial use by selected government agencies. These exceptions permit Claude’s use in legally authorized foreign intelligence analysis, such as combating human trafficking, identifying covert influence, and providing early warnings of military activities, facilitating preemptive diplomatic actions.
The policy allowing these exceptions is currently limited to AI models classified at AI Safety Level 2 (ASL-2) under Anthropic’s Responsible Scaling Policy. As the company progresses, it will continually assess its partnerships and impacts, ensuring AI serves the public interest and mitigates potential risks.
Anthropic has been dedicated to collaborating with governments to establish effective policies and testing regimes for AI. The company recently gave pre-release access to Claude 3.5 Sonnet to the UK Artificial Intelligence Safety Institute (UK AISI), which shared its findings with the US AI Safety Institute (US AISI). This partnership underscores Anthropic’s commitment to a safe transition toward transformative AI.
You can read the original article here: https://www.anthropic.com/news/expanding-access-to-claude-for-government
