E14 Public Audit

Inspiration

E14 Public Audit was inspired by the need for ordinary citizens to participate in public oversight without needing to be election lawyers, data engineers, or blockchain experts.

In Colombia, E14 election forms are public evidence, but reviewing them at scale is difficult. The information is fragmented, visual, and hard to compare manually. We wanted to make civic review more structured, transparent, and easier to preserve.

What it does

E14 Public Audit helps organize and review election form evidence so citizens and organizations can identify inconsistencies, preserve records, and build transparent review workflows.

The project is designed to support:

• E14 document intake
• OCR and visual review workflows
• structured inconsistency records
• evidence manifests
• public audit trails
• blockchain-oriented evidence anchoring
• citizen-friendly review batches

How we built it

The system combines data processing, document workflows, Supabase/PostgreSQL-backed records, and smart-contract evidence anchoring. It includes scripts and services for syncing public E14 data, preparing review batches, storing review results, rendering audit batches, and generating evidence artifacts.

How we used OpenAI and Codex

Codex helped us structure the workflow, reason about the evidence pipeline, document the public audit model, and prepare the project for a clear demo narrative.

OpenAI tools are useful here because civic data is messy: documents need to be transformed into structured review tasks, and the review process needs explanations that non-technical citizens can understand.

Challenges

The biggest challenge is trust. A civic audit tool must avoid overclaiming. It should help people inspect, compare, and preserve evidence, while being transparent about what is machine-assisted and what still requires human verification.

What we learned

AI can be powerful in civic technology when it is used to organize and explain evidence, not to replace public judgment. The most important feature is not automation alone, but a workflow people can audit.

What's next

Next we want to improve the reviewer interface, strengthen public evidence packaging, add better dashboards, and prepare pilot workflows for citizen groups and electoral transparency organizations.

Built With

• blockchain
• civic-tech
• codex
• foundry
• gpt-5.6
• javascript
• next.js
• ocr
• openai
• postgresql
• public-data
• python
• react
• solidity
• supabase
• typescript