Rules Foundation
ecosystem map
501(c)(3) — "Encode the law"
NSF POSE · Winter 2026
Arrow types
ideas / feedback
funding
technical / code
data / analysis
influence
Gov / Validation
Funding
Technical partners
Research
Contributors
Integrators
Validate statute encodings
against official interpretations
(transparency, auditability)
RLVR training signals
Policy reasoning benchmarks
(verifiable ground truth)
Computational law research
Legal formalization papers
(citations, datasets)
Fund public legal infra
Open-source civic tech
(measurable coverage)
Rules Foundation
Source document archive
Statutes, regulations → Akoma Ntoso XML
.rac DSL
Encoding format for law
Reference compiler
Open-source spec impl
AutoRAC validation harness
3-tier: CI tests → Oracle checks → LLM reviewers
Statute-linked test suites with citation chains
Ground truth test data
Open benchmarks for AI evaluation / RLVR
Archive · DSL · Validation · Benchmarks
All open source · Multi-stakeholder governance
5
Gov standards bodies
Tax Law Specialist, Reg. Counsel
(IRS, SSA, CMS, Treasury OTA)
Validate statute interpretations
Give: official guidance
3
AI labs
Research Scientist, Product Lead
(Anthropic, OpenAI, DeepMind)
Policy-aware AI, no hallucination
Give: compute, research collab
3
Academic researchers
Law Prof, Legal Informatics
(Stanford CodeX, MIT, law schools)
Computational law research
Give: validation, citations
2
Encoding community
Lawyer, Policy Analyst, Engineer
(Volunteers, pro bono, civic tech)
Review and improve encodings
Give: encoding review, interpretation
2
Downstream consumers
CTO, Product Lead, Eng Manager
(Cosilico, PolicyEngine, others)
Build on tested rule encodings
Give: feedback, bug reports
2
Grants / Foundations
Program Manager, Foundation Dir
(NSF, Sloan, Knight, Ford)
Fund public legal infrastructure
Give: grant funding
interpretations
audit trail
compute
ground truth
validation
datasets
encoding PRs
recognition
encodings
feedback
$ grants
impact reports
Assumptions to test
1. Agencies share interpretations?
2. AI labs invest in ground truth?
3. Multi-stakeholder governance?
4. Automation vs community?
Analogy
"OpenStreetMap for law"
Neutral, community-maintained
Scope: literal and bounded
Encode the law as written
Tooling/products → Cosilico
Downstream value chain
Cosilico: production compiler
PolicyEngine: policy analysis
AI agents: grounded reasoning
All consume RF encodings