California Expands State Agency Access to Anthropic's Claude AI

The California Department of Technology (CDT) announced an expanded partnership with Anthropic, granting state agencies broader access to the Claude family of large language models. This collaboration aims to furnish state employees with tools to enhance work efficiency and improve service delivery for Californians.

Under this agreement, California state agencies will gain access to Claude 3 Opus, Sonnet, and Haiku models, as well as the newly released Claude 3.5 Sonnet. Access will be provisioned through Anthropic's online chat interface. Additionally, the partnership facilitates access to Anthropic's Application Programming Interface (API), enabling agencies to integrate Claude models directly into their applications and workflows.

This initiative builds upon an existing collaboration established in September 2023, which initially provided a limited number of state agencies with access to Claude. The previous phase involved a pilot programme with state employees utilising the technology. Feedback from this initial deployment indicated that access to an advanced AI assistant could significantly improve job performance and accelerate task completion. For instance, employees reported using Claude for complex data analysis and summarisation, leading to faster operational outcomes.

The CDT underscored its commitment to responsible AI deployment. The department stated that all AI usage by state agencies must adhere to ethical guidelines and privacy protocols. The framework for AI adoption within state government prioritises security and the protection of sensitive information. This expanded access represents a tangible step in the state's broader strategy to integrate frontier AI applications within its operational infrastructure. The California Governor's Executive Order on AI, issued in September 2023, outlines a framework for responsible AI development and use across state government. This partnership aligns with that executive order's objectives.

As state entities increasingly integrate sophisticated AI, what new governance challenges will emerge regarding data privacy and algorithmic transparency?