ChatGPT Has Quietly Become the US Congress’s Favourite Paid AI Tool

The United States Congress is debating how artificial intelligence should be governed.
Behind the scenes, congressional staff are already using it.
ChatGPT accounted for approximately 90 per cent of recorded spending on standalone AI tools by House offices, committees and institutional accounts during the year ending 31 March 2026.
House expenditure records showed around $100,580 across 798 ChatGPT transactions, compared with at least $113,740 in total identified AI spending.
Anthropic’s Claude came second, receiving approximately $13,160 across 37 transactions. �
TechCrunch
What Is Congress Using AI For?
Congressional staff are reportedly using AI tools to:
Draft and summarise memoranda
Analyse legislation
Prepare hearing materials
Respond to constituents
Organise policy research
Draft social-media content
Process large volumes of documents
The recorded expenditure excludes free accounts and AI features bundled into broader software subscriptions, meaning actual adoption is probably higher than the payment records alone reveal. �
TechCrunch
The Political Spending Difference
Democratic offices accounted for approximately $54,165 in recorded AI purchases, more than three times the roughly $15,782 spent by Republican offices during the examined period. �
TechCrunch
That difference does not necessarily prove that Democratic staff use AI three times more.
Some offices may use free tools, centrally supplied software or AI features included in existing contracts.
Still, the figures show that artificial intelligence is becoming an ordinary congressional productivity tool rather than a distant policy concept.
Why the Story Went Viral
There is delicious political irony in the story.
Lawmakers are investigating AI safety, employment disruption, misinformation and corporate power—while their staff use the same technology to help analyse the legislation.
But that is not necessarily hypocrisy.
Governments must understand the tools they regulate.
The real issue is whether AI is being used with adequate safeguards.
The Risks Are Different Inside Government
A poorly drafted private email may embarrass one business.
An inaccurate congressional memorandum could influence legislation affecting millions of people.
Government AI use creates serious questions about:
Confidential information
Constituent privacy
Political bias
Hallucinated legal claims
Lobbying material presented as neutral analysis
Records retention
Cybersecurity
Public accountability
Staff must know whether information entered into a system is retained, used for model improvement or accessible outside the government’s security environment.
Can AI Improve Democracy?
Used carefully, AI could make government more responsive.
It may help staff:
Process constituent correspondence faster
Compare lengthy bills
Identify conflicting provisions
Translate public information
Summarise technical testimony
Make complex policy more accessible
But faster processing does not guarantee better judgement.
AI can organise public concerns without deciding which concerns deserve political priority.
That remains a human and democratic responsibility.
Mary Chuks’ Perspective
This is Human-in-the-Loop becoming government policy in practice.
AI can prepare the first summary.
A trained public servant must check the sources.
AI can identify patterns across constituent messages.
An elected representative must decide how to respond.
AI can draft legislation.
Human lawmakers must understand every provision they vote to impose on society.
The democratic danger is not government using AI.
It is government allowing AI-generated analysis to become invisible, unquestioned authority.
Conclusion
ChatGPT has become Congress’s leading paid AI tool because institutions eventually adopt technologies that save time—even while debating their risks.
The real test will not be how much Congress spends on artificial intelligence.
It will be whether lawmakers preserve human responsibility for the decisions AI helps them prepare.
Original Sources and Further Reading
Original reporting: TechCrunch, Rebecca Bellan, “Congress’ Favourite AI Tool? ChatGPT,” published 3 August 2026, citing House disbursement records and CNBC’s analysis. �
TechCrunch