Vitalik: Personal LLMs can be used to address attention issues in decentralized governance

On February 21, Ethereum founder Vitalik stated that artificial intelligence governance is dystopian: weak AI can lead to chaos, while powerful AI can result in catastrophic consequences. However, if AI is used properly, it can empower humans and promote the development of decentralized governance models. The core issue with decentralized governance models (including DAOs on Ethereum) is human attention limitations: thousands of decisions need to be made across numerous specialized fields, but most people neither have the time nor the ability to become experts in any of these areas, let alone master all of them. The usual solution is delegation, but this can weaken authority. Personal LLMs can be used to address attention issues, as follows: Personal Governance Agent: If governance mechanisms require users to make many decisions, personal agents can infer preferences based on personal writing, dialogue records, direct statements, etc., and represent the user in all necessary votes. Public Dialogue Agent: Good decision-making often cannot rely solely on linear processes, such as collecting opinions based on individual information and averaging (even quadratic averaging). A process is needed to aggregate information from many users, then give each person (or their LLM) the opportunity to respond based on this information. Market Proposal: If governance mechanisms value any form of “high-quality input” (which can be proposals or even arguments), a prediction market can be established. Anyone can submit input, and AI can bet tokens representing that input. If the mechanism “accepts” the input, it pays the token holder X dollars.