Vitalik proposes introducing AI-assisted DAO governance to address the issue of low voter participation.

Deep Tide TechFlow News, February 23 — According to Cointelegraph, Vitalik posted on the X platform on Sunday proposing to introduce personal AI large language models (LLMs) into decentralized autonomous organization (DAO) governance to address the long-standing issue of low participation rates in DAOs. Currently, the average voting participation rate in DAOs is only between 15% and 25%, which can lead to power concentration and pose governance attack risks.

Vitalik pointed out that the core challenge of democratized and decentralized governance is “the limited attention span of humans,” and that existing delegated voting mechanisms tend to allow a small number of representatives to control decision-making power, while other members almost lose their voice. He suggested that personal AI assistants could infer user preferences based on writing records, conversation history, and direct statements, and then cast votes on their behalf; if there is uncertainty about an important issue, the AI should proactively ask the user and provide relevant background information.

Regarding privacy protection, Vitalik proposed that personal LLMs could be placed in a “black box” environment to handle sensitive information, only outputting the final judgment results, thereby protecting privacy while supporting governance decisions involving confidential information.