A recent CNBC report, drawing from House disbursement records, highlights OpenAI's ChatGPT as the overwhelmingly preferred AI tool within the U.S. Congress. During the year ending March 31, OpenAI captured approximately 90% of all AI-related spending by House offices, committees, and institutional accounts. Out of a total of over $113,740 spent on AI, ChatGPT alone accounted for roughly $100,580 across 798 transactions, significantly outpacing Anthropic's Claude, which saw about $13,160 in spending.
Congressional staffers are leveraging these large language models for a diverse array of critical tasks. Their applications span from drafting internal memos and summarizing complex legislation to analyzing policy research, preparing materials for hearings, responding efficiently to constituent inquiries, and even crafting social media posts. This broad utility demonstrates the practical power of advanced natural language processing and generation in streamlining administrative and legislative workflows within a demanding, high-stakes environment.
For AI developers and researchers, this widespread adoption by a non-tech-native, high-profile institution like Congress is a significant validation. It underscores the maturity and practical utility of current LLM architectures, showcasing their ability to handle complex, sensitive, and varied real-world demands beyond typical tech use cases. This trend highlights the growing market for robust, user-friendly AI tools and provides valuable insights into the types of general-purpose AI capabilities that are most impactful, potentially guiding future research and development towards even more versatile and accessible applications.