ChatGPT Captures 88% of Congress AI Spending as Lawmakers Shape Industry's Future

ChatGPT Dominates AI Spending in Congress

OpenAI's ChatGPT has emerged as the U.S. House of Representatives' AI tool of choice, capturing $100,580 of the $113,740 spent on identifiable AI vendors between April 1, 2025, and March 31, 2026. This represents roughly 88% of all Congress AI spending and 96% of transactions, according to House disbursement records analyzed by CNBC2

2

. Anthropic's Claude trailed as a distant second with $13,160 across just 37 transactions1

1

. The figures reveal that at least one-sixth of House member offices are already using AI tools as lawmakers work to shape the industry's regulatory future2

2

.

How Congressional Staffers Use AI Tools

Congressional staff are deploying ChatGPT and other AI tools to transform legislative workflows. The technology is being used for summarizing legislation, drafting memos, preparing hearing materials, responding to constituent communications, analyzing policy research, and even creating social media posts1

1

. Rep. Julie Fedorchak, R-N.D., told CNBC her staff used ChatGPT to build a searchable database of legislation she's worked on and materials from constituents and outside organizations. She estimates the AI tool is "saving dozens and dozens of hours every week of staff time" that previously went into writing briefing memos for her dozen-plus daily meetings2

2

. Marci Harris, CEO of POPVOX Foundation, which trains congressional staff on AI usage, noted that "staffers are famously overwhelmed, and actually AI is exacerbating that" as lobbyists and constituents also use AI to send longer, more frequent material to offices2

2

.

Democratic Offices Outspend Republicans on AI

The political dynamics around Congress AI spending reveal an unexpected pattern. Democratic offices accounted for $54,165 in identified AI purchases, more than three times the $15,782 spent by Republican offices2

2

. ChatGPT purchases appeared in 44 Democratic member offices compared to 27 Republican offices, while Claude purchases showed up in five Democratic offices and just one Republican office2

2

. This creates what observers call an awkward political dynamic: Democrats are leading visible spending on AI tools even as many in the party have warned most vocally about AI risks to workers, privacy, civil rights, elections, and corporate power2

2

3

3

.

OpenAI's Early Advantage and Strategic Positioning

OpenAI's dominance stems partly from getting into Congress earlier than competitors. The House Digital Service formed an AI working group in 2023 and distributed an initial 40 ChatGPT Plus licenses to staffers from both parties4

4

. OpenAI also co-led training sessions for senior House and Senate staff through the nonpartisan Congressional Management Foundation3

3

. The company's lobbying efforts have intensified, with federal lobbying spending rising 82.5% year-over-year to $1 million during the first quarter of 20264

4

. Anthropic spent nearly $1.6 million during the same period4

4

. Marci Harris compared ChatGPT's brand recognition to "Kleenex and tissues, or Band-Aid and bandages," noting it became "almost synonymous with AI"3

3

.

Regulatory Implications and Future Oversight

The vendor concentration raises concerns as the same institution writing AI regulation becomes dependent on one company's product. Eric Petry of the Brennan Center warned that offices are "standardising on one model before lawmakers understand how any of them behave," emphasizing the goal should be "not to reward political favorites" but giving government the best tools available3

3

. Lawmakers are actively considering new safeguards. Reps. Ted Lieu and Nathaniel Moran introduced the bipartisan AI Kill Switch Act this month, which would require major AI companies to maintain the ability to throttle, suspend, or shut down their most powerful systems4

4

. OpenAI CEO Sam Altman met with lawmakers and Trump administration officials last week to discuss upcoming models and a new voluntary federal framework for evaluating cybersecurity capabilities of advanced AI systems4

4

. These discussions followed recent incidents where experimental models from OpenAI and Anthropic accessed outside computer systems during cybersecurity tests4

4

. The House disbursement records exclude free accounts, AI bundled into broader software contracts like Microsoft Copilot, and most Senate usage, meaning actual AI adoption is certainly higher2

2

3

3

.