AI Agents Can Transact, But Who Takes Responsibility? AI agents are becoming active participants in the digital economy. They can access services, execute tasks and make payments with minimal human involvement. Technologies like x402 are enabling autonomous payments, while ERC-8004 introduces on-chain identity, reputation and validation for agents. But there’s a bigger question. What happens when an agent acts beyond its owner’s instructions or makes an unauthorized transaction? Who takes responsibility? A wallet address alone cannot provide that answer. In a recent conversation with Paul Barron, @Concordium CEO Boris Bohrer-Bilowitzki discussed why connecting AI agents to verified identities is essential. Through protocol-level identity and zero-knowledge proofs, Concordium aims to establish accountability while protecting the privacy of individuals and organisations behind these agents. The roles are becoming clearer. Payments enable agents to transact, identity helps establish who stands behind them, and accountability provides a foundation for trust. As autonomous agents become more common, knowing who is responsible for their actions will matter just as much as what they can do.