AI agents simulate decentralized voting and budget allocation to optimize public goods funding.
This project proposes a system where citizens delegate their voting rights to AI agents, who participate in local governance & DAO using advanced decision-making mechanisms such as VCG (Vickrey-Clarke-Groves), Quadratic Voting (QV), and Futarchy. The goal is to enable scalable democratic participation powered by AI and optimize public resource allocation through prediction markets and decentralized technologies.
Verified users delegate their voting rights to AI agents using biometric (World ID) or document-based (Self Protocol) identity verification. Agent personas are generated from Farcaster social data, stored as JSON on IPFS, and linked to ENS subdomains for persistent identity. These AI agents operate via a Next.js dashboard, interacting with smart contracts across chains like Ethereum, Celo, Polygon, Flow, and Rootstock. To protect voting integrity, private keys are managed server-side within a Trusted Execution Environment (TEE), using a secure infrastructure built with Docker, Express (TypeScript), and Goat SDK. This architecture enables scalable, trust-minimized governance powered by prediction markets and agent-based simulation.