The US House spent at least $113,740 on AI tools in the year ending March, with OpenAI's ChatGPT accounting for about $100,580 across 798 transactions, making it the dominant tool for congressional offices.

Anthropic's Claude is a distant second, totaling around $13,160 across 37 transactions, far behind ChatGPT in spending.

The use of AI aims to boost efficiency in handling large volumes of information, but raises concerns about data security, accountability, and transparency in official processes.

Experts warn against bias toward a single vendor and urge access to multiple AI models to compare capabilities and inform regulatory decisions.

Lawmakers should evaluate several AI models to inform future regulations and ensure informed decision-making during the regulatory debate.

Legislative response is intensifying, including bipartisan efforts like the AI Kill Switch Act to require throttling or shutdown capabilities for powerful AI systems during emergencies.

The broader implication is that Congress's choice and procurement of AI tools could shape how AI is regulated and adopted across the public sector.

Congressional staff use AI tools like ChatGPT to write memos, summarize legislation, analyze bills, respond to constituents, prepare hearing materials, and draft social media posts, reflecting growing workflow automation.

ChatGPT's dominance signals a significant shift toward AI-assisted governance with ongoing implications for efficiency, democratic processes, and policy discussions on AI integration in public service.

The spending reflects competition among OpenAI, Anthropic, Google, and Microsoft to influence Washington policy and procurement as lawmakers shape AI purchasing and regulation.

OpenAI has extended influence through staff training partnerships, including collaboration with the Congressional Management Foundation to train staff on ChatGPT.

There is a partisan gap in AI spending: Democratic offices accounted for about $54,165, roughly three times the $15,782 spent by Republican offices, indicating differing approaches to AI adoption.