Investors say AI agent wallets are arriving faster than frameworks for liability and attribution.
• U.S. e-transactions law recognizes “electronic agents,” but modern autonomy raises harder questions about control and fault.
• OECD AI principles emphasize accountability and traceability by role, a model that maps to agent-wallet oversight.
Crypto investors are warning that AI agents with crypto wallets are moving from novelty to early deployment, creating legal and compliance risks before regulators and courts have clear rules on responsibility.
At a Feb. 24 panel at NEARCON 2026, Electric Capital partner Avichal Garg said developers are increasingly equipping autonomous agents with crypto wallets.
That, he argued, could allow software to hold assets, pay for services, trade tokens and even hire other AI agents.
The session, titled ‘The Checkbook of the Future: Who Holds the Keys?’, debated who controls capital in autonomous systems and what security, compliance and accountability require at scale.
The warning is not about whether agent wallets are possible. It is about what happens when they are common.
“AI itself cannot be punished,” Garg said, and there is still no clear answer on who bears responsibility if an agent with an independent wallet causes losses in transactions, lending or commerce.
Wallets turn autonomy into action. A model that can decide and pay can execute tasks continuously, at machine speed, across open networks. That is the appeal. It is also the risk.
In traditional finance, a customer is a person or a registered entity. Compliance programs are built around that assumption.
On-chain, a wallet can exist without a clear, public identity. If that wallet is controlled by an AI agent, the usual accountability hooks get slippery.
That creates a basic legal question: whose act is it when an agent signs a transaction that its operator did not specifically review?
“Electronic Agents” Exist, but the Fit Is Awkward
Automation is not new in law.
The Uniform Electronic Transactions Act (UETA) defines an “electronic agent” as a computer program or automated means used independently to initiate an action or respond to electronic records, without review or action by an individual at the time.
That helps establish that automated systems can form valid agreements. But it does not settle today’s crypto-native problems.
Modern AI agents can generate novel actions, operate in adversarial environments and interact with protocols that are not built around jurisdictional boundaries.