Steven Rouk

Project · July 2025

Reading a 130,000-word bill for its impact on farmed animals

When H.R. 1 (the “Big Beautiful Bill”) passed on July 4, 2025, I wondered what it meant for farmed animals. So I used AI to do a quick analysis.

The bill is 130,000+ words long, which is 70% as long as Harry Potter and the Goblet of Fire, making AI a helpful tool for “quick” analysis. I hadn’t seen any other policy analyses about the impacts of H.R. 1 on farmed animals, so I shared the results with other advocates in case they were helpful, both for the analysis itself and as a proof of concept for how AI might aid future policy analysis.

Words of bill text
132,743
Sections screened
46
Provisions flagged
126
Model cost (est.)
$0.88

How it worked

Since the bill is very long, I used a Python script to split it into 46 chunks along its titles and subtitles, and processed each chunk through OpenAI’s o3 model using the API, with a detailed prompt for how I wanted the analysis conducted: through the lens of farmed animal advocacy. The output was an executive summary of the biggest impacts (5 pages) and a full report with analysis for every section of the bill (59 pages). This process could be adapted fairly easily to analyze a bill through a different lens.

Where the animal impacts are

Each block is one section of the bill, in order. Width is the section’s length; color is how many provisions the AI flagged. The relevant material is small and scattered, which is why reading every section mattered.

30 of 45 sections of law had at least one flagged provision; 87 of the 126 flags were rated negative for animals. Hover or tap a block for details.
Show as a table
#SectionWordsFlaggedNeg.
1 Preamble and table of contents 3,453 0 0
2 Title I · Subtitle A: Nutrition 2,265 3 2
3 Title I · Subtitle B: Forestry 59 1 1
4 Title I · Subtitle C: Commodities 5,966 11 9
5 Title I · Subtitle D: Disaster Assistance Programs 925 6 6
6 Title I · Subtitle E: Crop Insurance 1,395 7 5
7 Title I · Subtitle F: Additional Investments in Rural America 1,809 12 8
8 Title II · Armed Services (general) 5,041 0 0
9 Title III · Banking, Housing, and Urban Affairs (general) 416 3 2
10 Title IV · Commerce, Science, and Transportation (general) 4,059 8 5
11 Title V · Subtitle A: Oil and Gas Leasing 3,287 4 3
12 Title V · Subtitle B: Mining 839 4 4
13 Title V · Subtitle C: Lands 2,241 6 3
14 Title V · Subtitle D: Energy 844 4 3
15 Title V · Subtitle E: Water 191 3 3
16 Title VI · Environment and Public Works (general) 1,273 6 5
17 Title VII · Chapter 1: Providing Permanent Tax Relief for Middle-Class Families and Workers 4,537 5 4
18 Title VII · Chapter 2: Delivering on Presidential Priorities to Provide New Middle- Class Tax Relief 7,983 0 0
19 Title VII · Subchapter A: Permanent U.S. Business Tax Reform and Boosting Domestic 9,740 6 4
20 Title VII · Subchapter A: Permanent Investments in Families and Children 15,457 5 5
21 Title VII · Subchapter A: Termination of Green New Deal Subsidies 12,291 5 3
22 Title VII · Chapter 6: Enhancing Deduction and Income Tax Credit Guardrails, and Other Reforms 2,642 2 0
23 Title VII · Subchapter A: Reducing Fraud and Improving Enrollment Processes 3,447 0 0
24 Title VII · Subchapter B: Preventing Wasteful Spending 1,173 0 0
25 Title VII · Subchapter C: Stopping Abusive Financing Practices 2,614 0 0
26 Title VII · Subchapter D: Increasing Personal Accountability 4,222 1 1
27 Title VII · Subchapter E: Expanding Access to Care 1,075 0 0
28 Title VII · Subchapter A: Strengthening Eligibility Requirements 798 0 0
29 Title VII · Subchapter A: Improving Eligibility Criteria 2,161 0 0
30 Title VII · Chapter 4: Protecting Rural Hospitals and Providers 2,120 0 0
31 Title VII · Subtitle C: Increase in Debt Limit 44 0 0
32 Title VII · Subtitle D: Unemployment 336 0 0
33 Title VIII · Subtitle A: Exemption of Certain Assets 179 2 1
34 Title VIII · Subtitle B: Loan Limits 1,294 3 2
35 Title VIII · Subtitle C: Loan Repayment 5,117 3 0
36 Title VIII · Subtitle D: Pell Grants 1,556 3 1
37 Title VIII · Subtitle E: Accountability 950 0 0
38 Title VIII · Subtitle F: Regulatory Relief 334 0 0
39 Title VIII · Subtitle G: Garden of Heroes 100 0 0
40 Title VIII · Subtitle H: Office of Refugee Resettlement 456 1 0
41 Title IX · Subtitle A: Homeland Security Provisions 1,895 4 2
42 Title IX · Subtitle B: Governmental Affairs Provisions 822 3 0
43 Title X · the Judiciary (general) 9,222 1 1
44 Title X · Immigration and law enforcement: appropriations 3,268 2 2
45 Title X · Subtitle B: Judiciary Matters 304 2 2
46 Title X · Subtitle C: Radiation Exposure Compensation Matters 2,543 0 0

What it flagged

Four of the findings the AI rated most significant, checked against the bill text in October 2026:

  • Manure fuel credit §70521. The rewritten clean fuel credit bars any fuel from getting an emissions rate below zero, with one exception: fuel derived from animal manure. The AI read this as a new revenue stream for large dairy and hog operations.
  • Poultry insurance pilot §10507. A new federally backed insurance product covering contract broiler and egg growers’ weather-driven utility costs, with no welfare conditions attached.
  • Paid fast-track environmental reviews §60026. Project sponsors can pay a fee for expedited NEPA reviews, which the AI flagged as a way to speed factory-farm permitting.
  • Western water infrastructure §50501. $1 billion for water conveyance and storage. The link to feed crops and dairy is the AI’s inference.

What this is for, and what it isn’t

I did not do thorough validation of the analysis (I read and spot-checked some things), so it’s possible that large portions of it are incorrect. All statements and conclusions should be independently verified before use. That goes double for specific numbers, like the summary’s estimate of roughly $80 billion toward animal agriculture.

I think this approach could be very good for quickly discovering insights, keeping up with things at a pace that’s otherwise impossible, and conducting “good enough” analyses when you don’t have the expertise to do so. Rather than this being the last step of research, I would view it as a first step, with the experts picking it up from here.

If AI analysis is good enough and accurate enough, this type of approach could be used to automatically track and analyze bills for their impacts on farmed animals, so advocates can respond more quickly. Any publicly available text on the internet can now be analyzed by AI for animal impacts. I think there’s probably a lot of opportunity there.

Read the reports