Ethereum is set to enhance its capabilities in the AI sector with the introduction of EIP-8004, titled “Trustless Agents,” targeting implementation in the anticipated 2026 hard forks like Gloas-Amsterdam and Heze-Bogota. This Ethereum Improvement Proposal aims to establish the blockchain as the primary platform for autonomous AI agents by integrating a trustless interaction framework. ERC-8004 is designed as a universal trust layer that enables AI agents to transact and collaborate securely across organizational boundaries without requiring pre-existing trust relationships.

The standard proposes three on-chain registries—Identity, Reputation, and Validation—that help AI agents discover one another, maintain verifiable reputations, and validate their autonomous actions cryptographically and economically. Agent identities will be represented via ERC-721 tokens, making them compatible with popular wallets like MetaMask and Trust Wallet. This infrastructure supports a vibrant agentic economy where autonomous agents can perform diverse tasks such as customer service or supply chain management while maintaining transparent performance records on-chain.

ERC-8004 is expected to complement existing agent communication protocols, like Agent-to-Agent (A2A), by providing the necessary blockchain-based discovery and verification services. By enabling open, decentralized AI agent markets free of centralized gatekeepers, the proposal fosters innovation and security in autonomous interactions. Ethereum’s adoption of this protocol reflects its broader AI strategy, positioning itself at the forefront of decentralized AI ecosystem development and autonomous software agent economy growth.

Frequently asked questions

What is EIP-8004?

EIP-8004, titled 'Trustless Agents,' is an Ethereum Improvement Proposal aiming to enhance the blockchain for autonomous AI agents.

How does ERC-8004 support AI agents?

ERC-8004 establishes a universal trust layer that enables secure transactions and collaborations for AI agents across organizational boundaries.

What are the key components of the proposal?

The proposal includes three on-chain registries—Identity, Reputation, and Validation—to help AI agents discover and validate each other.