A Hybrid Blockchain–AI Architecture for Secure and Transparent Decentralized Crowdfunding

Abstract

Crowdfunding websites tend to have centralized escrow infrastructure, which can create concerns over the lack of transparency, security threats, and fraud vulnerability. The proposed system, a hybrid blockchain–AI architecture, combines Ethereum-based smart contracts and machine learning-based fraud detection to result in a decentralized and transparent crowdfunding space. The blockchain layer ensures accountability with controlled release of funds based on milestones, limiting the tendency to spend funds more due to the cryptocurrency nature with the AI module detecting fraudulent activity based on analysis of textual, transactional, temporal, and reputation data. Experimentation proves that the proposed system promotes higher trust, reduces transaction cost, and signifies a robust fraud detection model compared to conventional crowdfunding models. The results suggest creating a combination of the immutability of blockchains with the analytical power of AI as a potential route to safer and more effective decentralized finance apps.