Unbiased Redistricting Algorithm, Built With AI on a Leash
Christopher Melson built Fair Maps, an unbiased redistricting algorithm, and governed the AI coding assistant that helped write it. The business risk is specific. When public trust is the product, one unrecorded tie-break, one untested change or one unreviewed AI edit can quietly tilt a map and end the credibility of the whole project. His decision was to pay for mechanical enforcement instead of relying on reviewer vigilance. Every change starts as a written spec and a plan, gets an independent review, and then faces continuous integration checks that fail the build instead of warning. A publish step refuses to replace any map that changed without a version bump. The costs are real: setup time, long runs, no tuning. The evidence is concrete. Version 1.0.0 was cut on 2026-10-11, and the legislative release then added State House and State Senate maps, 7,210 districts in all, on a drawing engine whose code did not change. The next release, engine 2.0.0, fixed two bugs in how the rules were applied and took a major version bump. Leaders weighing AI-written code where neutrality matters can copy the pattern: make the rules checkable by anyone, then make the checks impossible to skip.
- Core Argument
- An output nobody can steer comes from deleting choices (ties, caps, rounding, seeds, engine math), not from supervising them. The same logic governs an AI coding assistant: nothing merges without gates a human designed.
- Target Audience
- Operators, engineering leaders and civic-tech builders deciding how to govern AI-written code where public trust is the product.
- Key Insight
- At the legislative release, the unchanged 1.0 engine drew 7,210 districts, and fixtures, a publish gate and SHA-256 fingerprints made that claim checkable by anyone.