Anthropic and UK Government Partnership for GOV.UK AI Assistant

Anthropic and the UK's Department for Science, Innovation and Technology (DSIT) have partnered to build and pilot an AI-powered assistant for GOV.UK. This initiative aims to help citizens navigate government services and receive tailored advice, starting with a primary focus on employment services to assist users in finding work, accessing training, and understanding available resources.

Claude-Powered Agentic Support for Public Services

The GOV.UK AI assistant is an agentic system powered by Claude, designed to move beyond simple question-answering to actively guide users through government processes with personalized support. This project is a direct outcome of a Memorandum of Understanding signed in February 2025 between Anthropic and the UK government to explore the transformation of public services through advanced AI.

To ensure long-term sustainability, Anthropic engineers are collaborating with civil servants and software developers at the Government Digital Service. This partnership is designed to build AI and AI safety expertise within the UK government, enabling the government to independently maintain the system over time.

Initial Use Case: Employment and Workforce Re-entry

The AI assistant's initial rollout focuses on supporting job seekers entering or re-entering the workforce. Key capabilities include:

• Personalized Guidance: Providing tailored career advice and routing users to the correct services based on their specific circumstances.
• Resource Access: Helping users find and access training and explaining available supports.
• Contextual Memory: Maintaining context across interactions so users do not have to repeat information in subsequent sessions.
• Data Privacy: Users maintain full control over their data, including the ability to opt out or manage what the system remembers, with all personal information handled according to UK data protection law.

Implementation Framework and Safety Standards

The project follows DSIT's "Scan, Pilot, Scale" framework, a phased approach that allows for iterative testing and learning before a wider rollout.

Safety is a core component of the deployment, with Anthropic continuing to work with the UK AI Security Institute to test and evaluate models. The evaluation frameworks developed through this collaboration inform how Claude is deployed within the public sector to ensure the system remains safe and trustworthy.

Global Context of Public Sector AI Deployment

This partnership is part of a broader strategy by Anthropic to deploy AI for public benefit globally. Other similar initiatives include:

• United Kingdom: Partnering with the London School of Economics to provide students access to Claude.
• Iceland: Collaborating with the Ministry of Education and Children to launch a national AI education pilot for teachers.
• Rwanda: Partnering with the Rwandan Government to bring AI education to hundreds of thousands of learners.