House Bill Seeks Rules for AI Agent Transactions, Keeps Humans and Businesses Accountable

A proposed House measure would establish rules for transactions performed by artificial intelligence agents, while keeping legal responsibility with the people and businesses that authorize them.

House Bill No. 11014, or the proposed “AI Agent Transactions and Accountability Act,” was filed by Camarines Sur Representatives Migz Villafuerte and Luigi Villafuerte.

The measure addresses AI systems authorized to negotiate, execute contracts and electronic documents, initiate transactions, and interact with application programming interfaces, smart contracts, and other automated technologies.

Its central principle is that an AI agent would not become a separate legal person. An authorized act would instead be attributed to its principal—the person or entity behind it.

The bill’s text states that using an AI agent would not, by itself, invalidate an otherwise lawful electronic contract, signature, payment instruction, or other electronic act. It would not give the agent independent contractual capacity or legal personality.

Under the proposal, the Department of Information and Communications Technology (DICT) would establish a voluntary registration system and Digital Authority Credential framework.

The credential would help third parties verify a principal’s identity and the scope of authority delegated to an agent without requiring access to secret authentication information.

Mandatory registration would be limited to narrowly defined high-value, high-risk, or systemically significant autonomous transactions where existing regulations are inadequate. The proposal would therefore not require registration for every AI agent or ordinary use of AI.

The measure also includes protections for lawful digital asset ownership, self-custody, peer-to-peer transactions, and open-source technologies. It would not require disclosure of private keys, seed phrases, or passwords.

Principals would generally remain responsible for acts legally attributable to their agents. The proposal identifies exceptions involving circumstances such as stolen or compromised credentials, unauthorized access, fraud, malicious interference, and certain failures beyond the principal’s control.

It would also protect good-faith third parties relying on valid Digital Authority Credentials. Registered agents performing legally consequential autonomous transactions would need to maintain proportionate transaction records.

For specified knowing or willful offenses involving fraud or deliberate circumvention of applicable law, the bill proposes fines of ₱200,000 to ₱2 million, imprisonment of six months to six years, or both, without prejudice to liability under other laws.

These are proposed penalties for serious misconduct, rather than a blanket punishment for errors made by an AI system.

The Villafuerte lawmakers also authored HB No. 2827, the proposed “Artificial Intelligence Development and Regulation Act,” which seeks a broader framework for AI development, deployment, regulation, and oversight.

Together, the measures address different aspects of expanding AI use: the wider governance of the technology and the specific question of who answers for transactions carried out through automated agents.

HB 11014 has been filed in the House of Representatives and has not been enacted into law.

Find this story on GoogleHouse Bill Seeks Rules for AI Agent Transactions, Keeps Humans and Businesses Accountablegoogle.com · Opens in a new tab↗