The Near Foundation is actively developing AI-powered "digital twins" to participate in governance votes within its decentralized autonomous organization (DAO). This initiative aims to address the prevalent issue of low voter participation in many protocols.
Lane Rettig, a researcher at the Near Foundation specializing in AI and governance, revealed that this AI-driven governance overhaul is currently under development. The core concept revolves around creating a "digital twin" for each user, capable of learning their preferences and making governance decisions on their behalf. This approach seeks to transform the voting process into a streamlined "math problem" that can be executed almost instantaneously.
Addressing Voter Apathy
The Near Foundation's initiative directly targets the problem of voter apathy within DAOs. Average participation rates in DAOs typically range between 15% and 25%. This low engagement can lead to centralization of power, ineffective decision-making, and increased vulnerability to governance attacks. In such attacks, malicious actors can exploit low turnout to pass damaging proposals without sufficient opposition.
Rettig explained the envisioned end-game: "That's in a sense, almost our end game vision for this, where we replace all human actors with a digital twin, if you want to call it that, to solve this voter apathy, participation issue". The AI delegate would learn user preferences regarding projects and fund allocation and then act accordingly. The AI would proactively inform users of relevant proposals.
Human Oversight and Gradual Rollout
While the project aims for significant automation, human input will remain an essential component. Rettig emphasized the importance of keeping "a human in the loop," particularly for critical decisions such as fund allocations or strategy pivots. He believes certain proposals are too important to be left solely to AI decision-making.
The rollout of AI delegates will be a phased process. Early models will function as chatbots, offering advice and information on proposals to better inform users. The AI can also fill out basic templates. The deployment will start with AI delegates representing large groups with similar voting preferences before transitioning to individual delegates and potentially even AI delegate CEOs.
Ensuring Alignment and Verifiability
To ensure AI delegates remain aligned with user values, the Near Foundation is implementing a verifiable model training model. This model will provide cryptographic proof of the AI's training cycles and inputs, enhancing transparency and trust.
AI in Governance: Broader Trends and Concerns
Investment manager VanEck estimates the number of AI agents in the crypto industry surpassed 10,000 by the end of 2024 and expects it to exceed 1 million by 2025. While AI offers potential benefits, concerns exist regarding security risks and the potential for errors in critical decisions if AI is relied upon too heavily.
The Near Digital Collective, Near's main DAO, has already implemented an AI tool called Pulse. Pulse tracks community sentiment, summarizes Discord forums, and highlights important content. House of Stake (HoS) is envisioned as a community-driven governance framework for NEAR, supported by the NEAR Foundation and meant to eventually operate independently. HoS plans to leverage AI tools as part of its governance suite.
Ultimately, the Near Foundation's work reflects a growing trend toward AI-enhanced governance in Web3. DAOs are exploring various AI integrations, from AI assistants that summarize forum debates to autonomous agents that can vote on proposals. The goal is to create more efficient, data-driven governance processes while addressing the challenges of low participation and ensuring community alignment.
