Ethereum Proposal Targets Verifiable AI Identity
A new Ethereum proposal, ERC 8004, is under development to create a verifiable identity and reputation system for AI agents. The standard aims to establish a decentralized trust layer for interactions as AI agents become more autonomous. This system could become a foundational security and compliance layer for future platforms where AI agents transact with each other.
- The proposal, officially titled "Trustless Agents," was a collaborative effort co-authored by representatives from MetaMask, the Ethereum Foundation, Google, and Coinbase. - ERC-8004 establishes three on-chain registries: an Identity Registry using ERC-721 for portable agent identifiers, a Reputation Registry for standardized feedback, and a Validation Registry for cryptographic and economic verification of an agent's work. - The standard extends Google's existing Agent-to-Agent (A2A) protocol, adding a layer of trust to enable interactions in open, decentralized environments. - After its official proposal on August 13, 2025, and a period of community review, ERC-8004 went live on the Ethereum mainnet on January 29, 2026. - The system is designed to be modular and can be deployed on the Ethereum mainnet or more cost-effective Layer 2 networks like Optimism or Arbitrum. - To ensure auditability, on-chain pointers and hashes within the system cannot be deleted, creating a permanent record of agent interactions and reputation. - The protocol addresses potential Sybil attacks, where fake agents could inflate their reputation, by making all feedback signals public under a consistent schema. - Smart contracts can programmatically interact with AI agents based on their on-chain trust scores, enabling automated workflows such as escrow releases that are contingent on validated task completion.
