Through public records requests, I obtained thousands of pages of ChatGPT conversation logs from local government officials in Washington state. The records showed that public servants have quietly used the AI chatbot for a wide variety of tasks — including generating policy documents, talking points, comprehensive plan updates, responses to audit recommendations, social media posts, replies to constituent emails and more.
The volume of records indicated widespread usage of the technology, and formed the basis for a two-part investigation published in August 2025 by Cascade PBS and KNKX. While some records showed the potential utility of generative AI, others raised questions about transparency, authorship, security and bias. While state guidance indicates AI-generated government documents be labeled as such, none of the records reviewed in this reporting included disclosures that AI was used in their production.
The investigation also found that, as generative AI tools are being used more and more by local governments, adoption is often outpacing policies and safeguards for ethical use.
