The Harder Problem Action Fund is an advocacy organization fighting harmful AI consciousness legislation. We track pending bills, score legislation, lobby for evidence-based policy, and mobilize public action before ignorance becomes law.
Legislative Watch
Methodology
Our approach to monitoring legislation and assessing legislators. This document explains what we measure, how we score, and the limitations of our assessments.
Overview
What These Assessments Are
- Impact-based assessments of legislation affecting AI consciousness policy
- Voting record compilations for legislators with relevant positions
- An advocacy tool informed by transparent methodology
- A resource for constituents engaging with representatives
What These Assessments Are Not
- A prediction of if or when AI sentience will emerge
- A determination of whether any current AI is sentient
- A comprehensive evaluation of a legislator's overall record
- A position on the merit of legislators as individuals
Purpose
The Legislation Tracker and Legislator Scorecards serve the Action Fund's mission of influencing policy outcomes related to AI consciousness. Unlike our sister organization's educational Sentience Readiness Index, these tools are explicitly designed for advocacy. We assess whether legislation helps or harms the policy environment for evidence-based AI consciousness policy, and whether legislators advance or impede that goal.
Organizational Note
The Harder Problem Action Fund is an advocacy organization. Unlike educational nonprofits, the Action Fund may engage in lobbying and political activity. This methodology reflects our advocacy mission while maintaining rigorous analytical standards.
What We Measure
Core Questions
For Legislation
"What impact would this legislation have on the policy environment for AI consciousness research, recognition, and response?"
For Legislators
"Based on their voting record and public positions, how does this legislator's cumulative record affect AI consciousness policy?"
Key Concepts
Threat
Legislation that would foreclose future options, restrict research, or entrench positions before scientific questions are resolved. High-threat bills are difficult to reverse and set negative precedents.
Opportunity
Legislation that preserves flexibility, enables research, or creates mechanisms for adaptive response. Opportunity bills open doors rather than close them.
Voting Record
The complete pattern of a legislator's votes, co-sponsorships, public statements, and committee actions on AI consciousness-related matters.
Bill Impact Scoring Framework
How Bill Scoring Works
Every bill receives two assessments:
1. Direction
Is this bill harmful or helpful to AI consciousness policy?
2. Impact Score (1–10)
How significant is this bill? Uses the four weighted categories below.
Examples:
• OH SB 847: Threat Impact 9.5 — Very
harmful bill we strongly oppose
• CO AI Act: Opportunity Impact 8.0 —
Very helpful bill we strongly support
Each bill's Impact Score is derived from four weighted categories (below). The Direction classification is based on our assessment of whether the bill helps or harms AI consciousness policy.
Category A: Scope of Impact (30%)
How broad and deep are the effects if this legislation is enacted?
A1. Jurisdictional Reach (0-25 points)
Does the bill affect local, state/provincial, national, or international policy? Higher-level jurisdictions receive higher impact scores due to broader reach.
A2. Domain Breadth (0-25 points)
Does the bill affect a single domain (e.g., research funding) or multiple domains (research, legal status, institutional engagement, professional practice)?
A3. Population Affected (0-25 points)
How many people, institutions, or entities would be directly affected by implementation? Weighted by jurisdiction population and institutional density.
A4. Preemption Effects (0-25 points)
Does the bill preempt lower-jurisdiction action? State bills that preempt local ordinances or federal bills that preempt state action score higher.
Category B: Reversibility (25%)
How difficult would it be to undo this legislation if circumstances change?
B1. Legal Mechanism (0-33 points)
Is this a statute, regulation, constitutional amendment, or court decision? Constitutional amendments are nearly irreversible; regulations are more easily changed.
B2. Sunset Provisions (0-33 points)
Does the bill include automatic expiration, mandatory review periods, or mechanisms for reconsideration? Absence of sunset clauses increases threat score.
B3. Political Difficulty (0-34 points)
What political conditions would be required to reverse this legislation? Bills requiring supermajorities or ballot initiatives to reverse score higher.
Category C: Precedent Risk (25%)
How likely is this bill to inspire similar legislation in other jurisdictions?
C1. Template Potential (0-33 points)
Is this bill designed as model legislation? Does it use language that could be easily adopted by other jurisdictions? Evidence of coordination increases score.
C2. Jurisdiction Influence (0-33 points)
How influential is this jurisdiction in setting policy trends? Early-mover states (California, Texas, New York) or countries (UK, Canada) score higher.
C3. Cross-Border Applicability (0-34 points)
Could this bill's framework be adopted internationally? Bills creating novel legal concepts or definitions that could spread across borders score higher.
Category D: Likelihood of Passage (20%)
What is the probability this bill becomes law?
D1. Sponsor Strength (0-25 points)
Political capital, seniority, and track record of the bill's sponsors. Committee chairs and majority leaders increase passage likelihood.
D2. Committee Composition (0-25 points)
Ideological and partisan composition of the relevant committee(s). Favorable committee alignment increases score.
D3. Floor Vote Prospects (0-25 points)
Based on party control, whip counts (where available), and analogous vote patterns, what is the likely outcome of a floor vote?
D4. Executive Disposition (0-25 points)
Is the relevant executive (governor, prime minister, president) likely to sign or veto? Veto-proof majorities score higher regardless.
Impact Level Classification
Impact scores translate to urgency levels for both Threats and Opportunities:
For Threat Bills:
For Opportunity Bills:
Legislator Scoring Framework
Legislators are assessed on a 0-100 point scale based on their cumulative record on AI consciousness-related matters. The score converts to a letter grade.
Scored Actions
Floor Votes (±10 points each)
Recorded votes on AI consciousness-related legislation. Positive votes (opposing harmful bills or supporting beneficial ones) add points; negative votes subtract. Abstentions score 0.
Co-Sponsorship (±5 points each)
Lending name and political capital to legislation. Primary sponsorship of harmful bills scores −10; primary sponsorship of beneficial bills scores +10.
Committee Actions (±3 points each)
Voting in committee, proposing amendments, scheduling or blocking hearings. Committee chairs receive double weight for procedural actions.
Public Statements (±2 points each)
Official statements, floor speeches, press releases, and verified social media posts taking positions on AI consciousness policy. Op-eds in major outlets score ±3.
Leadership (+5 points)
Taking visible leadership roles: convening hearings, organizing coalitions, leading floor debate, or serving as a public champion for evidence-based policy.
Constituent Responsiveness (±2 points)
Based on reported constituent interactions: responsiveness to constituent concerns, willingness to meet, quality of engagement on the issue.
Grade Classification
Legislators start with a baseline score of 50 (neutral). Actions move them up or down.
Connection to Legislation Tracker
Legislator scores are derived from actions on bills tracked in our Legislation Tracker. The "Key Votes" column in the scorecard table shows specific votes on tracked legislation. This creates a coherent data model: the same bills appear in both tools, and legislator grades reflect their actions on those specific bills.
Data Sources
Primary Sources
- USA: Congress.gov, OpenStates API, state legislature websites
- UK: Parliament.uk, TheyWorkForYou API, Hansard
- Canada: OpenParliament.ca, LEGISinfo, provincial legislature sites
- Official government gazettes and registries
- Recorded floor votes and committee transcripts
- Official press releases and legislator websites
Supplementary Sources
- Major news outlets for reporting on legislative developments
- Verified social media accounts for public statements
- Think tank and policy organization analysis
- Academic research on legislative behavior
- Constituent reports (used cautiously, verified where possible)
Source Reliability Hierarchy
Assessment Process
Data Collection
Monitor legislative databases for new and updated bills. Gather voting records, co-sponsorships, and public statements.
Initial Assessment
Staff analyst applies scoring rubrics to generate initial threat scores and legislator points using standardized criteria.
Review & Publication
Senior editor reviews for consistency, accuracy, and methodology compliance. Assessment published with full sourcing.
Update Cycle
Bill Status
Updated within 48 hours of major legislative action
Threat Scores
Reassessed weekly or as conditions change
Legislator Scores
Updated after each scored action, reviewed quarterly
Limitations & Caveats
Methodological Limitations
- Novel policy area with limited historical precedent for scoring
- Subjective elements remain despite structured rubrics
- Data availability varies across jurisdictions
- Threat score "likelihood" estimates are inherently uncertain
- Legislator records may be incomplete for newer or less active members
- AI-related legislation may have consciousness implications not explicitly stated
Scope Limitations
- Only assesses legislation explicitly or implicitly affecting AI consciousness policy
- Does not assess legislators' overall quality or fitness for office
- Coverage limited to USA, UK, and Canada (expanding)
- May miss local ordinances or regulations not in primary databases
- Cannot capture informal political dynamics or behind-the-scenes negotiations
Interpretation Guidance
- These are advocacy tools, not neutral assessments. We have a position: legislation that forecloses options before scientific questions are resolved is harmful. Scores reflect this position.
- Threat scores are not predictions. A "Critical" bill may fail; a "Low" bill may pass. Scores reflect potential impact weighted by probability, not certainty.
- Legislator grades are narrow. A low grade means a problematic record on this specific issue, not that the legislator is "bad" overall.
- Use multiple sources. These tools are one input among many. Consult primary sources, other watchdog organizations, and your own judgment.
Changelog
Initial methodology release. Established four-category bill threat scoring framework and six-action legislator scoring system. Coverage: USA, UK, Canada.
Methodology updates will be documented here with version numbers and effective dates.
Explore the Data
Now that you understand our methodology, explore the legislation we're tracking and see how your representatives score.