Industry

ChatGPT Received Majority of Congress’s Reported AI Spending, CNBC Finds

CNBC analysis of House disbursement records shows OpenAI’s ChatGPT accounted for about $100,580 of at least $113,740 in AI-related purchases by House offices, committees and institutional accounts in the year ending March 31.

ChatGPT Received Majority of Congress’s Reported AI Spending, CNBC Finds

Analysis of House disbursement records by CNBC shows that OpenAI’s ChatGPT received the largest share of reported AI-related purchases by House offices, committees and institutional accounts in the year ending March 31, 2024.

  • CNBC found total reported AI spending of at least $113,740, of which about $100,580 (across 798 transactions) was spent on ChatGPT.
  • That places OpenAI at roughly 90% of the reported AI expenditures.
  • Anthropic’s Claude ranked second with $13,160 spent across 37 transactions.

The CNBC analysis notes that these figures exclude any AI use through free accounts and AI functionality that is bundled into broader software contracts, so actual usage and spending could be larger than the recorded amounts.

How congressional staff are using AI

According to the report, congressional staff have been applying ChatGPT and other tools for a range of tasks, including:

  • drafting memos and background briefs;
  • summarizing and analyzing legislation;
  • responding to constituents;
  • preparing materials for hearings;
  • sorting and synthesizing policy research;
  • composing social media posts.

Partisan breakdown

The data indicate that Democratic offices accounted for $54,165 in reported AI purchases during the period, roughly three times the $15,782 recorded for Republican offices.

Why the numbers matter

The findings highlight increasing adoption of AI tools within Congress and show that a single provider, OpenAI, dominates the reported expenditures. At the same time, exclusions in the dataset—such as free accounts and bundled services—mean the recorded totals represent a partial view of congressional AI use.