Buterin: Human Attention Limits Hinder DAOs; Proposes Personal AI Stewards

Vitalik Buterin identified limits to human attention as the fundamental problem facing decentralized autonomous organizations (DAOs) and democratic governance, arguing on X that participants confront thousands of decisions across diverse domains without sufficient time or expertise to evaluate them all ^[1][3].

To address that attention bottleneck, Buterin proposed a governance model in which personal AI agents act as stewards that learn users' preferences from their writing, history and stated values and vote or recommend decisions on their behalf, automating routine participation while remaining aligned with individual preferences ^[2][3].

He and coverage of the proposal emphasize technical safeguards — including zero‑knowledge proofs, multi‑party computation (MPC) and trusted execution environments (TEEs) — intended to protect voter identity, prevent coercion or bribery, and preserve privacy while enabling AI‑assisted participation ^[4][2].

Buterin framed the approach as a way to scale democratic participation rather than centralize control, arguing that correctly designed AI stewards and cryptographic protections could help DAOs overcome the human attention constraint ^[3][4].