zXdao

What It Solves

As college students steering the university DAO, we have identified critical challenges hindering optimal DAO functionality:

• Anonymity Gap: Existing DAOs lack a mechanism for voter anonymity, compromising user privacy.
• Quorum Deficiency: High instances of proposals failing to reach quorum hinder effective decision-making.

Supported by findings from research, revealing a meager 5% average participation per proposal, these issues underscored the need for innovation.

In response, we proudly present zXdao, pioneering a transformative approach to unlock the full potential of DAOs.

Our Solutions

• AI-Driven Participation: Leveraging VADER and traditional Machine Learning, we construct a behavioral model of each user through sentiment analysis. In cases where users fail to vote by the proposal deadline, our AI acts on their behalf, ensuring continuous and representative participation.
• zk Anonymity: Employing cutting-edge zk proofs and the Semaphore protocol, we introduce robust anonymity measures. Users validate their identity commitment by being part of the DAO. This process empowers users or their corresponding AI to vote securely from any wallet, supported by a comprehensive zk proof.

zXdao Impact

• Enhanced Participation: By addressing the anonymity concern, zXdao fosters a more inclusive environment, encouraging higher engagement and participation.
• Quorum Assurance: AI-driven voting mitigates the risk of proposals failing to achieve quorum, promoting effective decision-making within the DAO.

Links

• https://github.com/gauthking/ethindia-2023-proj

Addressing AI challenges:

• Crafting an accurate user behavior model with minimal input was key.
• Exploring sentiment analysis tools like BERT and LSTMs, VADER emerged as the optimal choice, tailored for social media inputs with interpretable features.
• Predicting votes on new proposals required meticulous feature selection from sentiment scores for the ML model.
• Future plans involve continuous learning, using new proposals as training data, and integrating explainable AI for transparency in the voting process.

Addressing Anonymity challenges:

• We initially encountered double signaling with other protocols. Shifting to Semaphore resolved this issue.
• Storing group instances on-chain wasn't supported by the contract, prompting the creation of a dedicated off-chain database for managing group IDs.
• In instances where Semaphore wasn't available on all networks like Scroll, Base, and zkEVM, we proactively deployed our Semaphore contracts manually, ensuring consistent functionality across diverse environments.