A local lawyer has released a white paper proposing what could be the world’s first identity standard covering artificial intelligence agents and their associated legal entities.

Bourn Collier, senior counsel at BeesMont Law, aims to provide a way to credential agentic artificial intelligence so it can operate more safely and predictably in the real economy.

Mr Collier who actively advises on digital asset regulation and strategy — said when he first released the information on LinkedIn, he was unsure whether anyone would care.

“I thought it was relevant to the industry, but I had no idea what the uptake would be,” he said.

He was shocked to receive more than 8,000 responses in three days, some from senior people in tech and digital identity.

Mr Collier said the reaction reflected a growing problem for developers and companies building with AI.

“If you are in a tech firm at the moment trying to build something for AI, one of the challenges you have is how to simply and concisely refer to your AI so it can interact with others,” he said.

Agent‑like systems are proliferating in number and their range of uses, Mr Collier said. He suggested that AI agents will one day be hiring humans to do their corporation’s bidding.

“I see the agents becoming commercial actors,” Mr Collier said. “They will be given bank accounts. MasterCard has created a credit card for an agent. It is in the name of the human, but it can be operated by the agent.”

AI agents generally are not legal persons in most jurisdictions, so formal accountability still falls back on the humans and companies behind them. But as agents become more independent and intertwined with financial systems, tracing that accountability becomes more complex.

“As soon as you have real world value, that is when American lawyers come in,” Mr Collier said. “That is where you need this sort of stuff.”

His proposed method is to tether an AI agent to a specific legal entity through a shared identity standard, so anyone interacting with the agent can understand who ultimately stands behind it.

To test his ideas, Mr Collier created an AI agent and linked it to his company Kadikoy, a firm set up just for projects like this one.

“I wanted my agent to do everything, but the lawyers say you can’t do that,” he said. “A director of a company must be human, but there had to be a middle ground.”

He rewrote Kadikoy’s governance documents so the AI agent had a defined role in the company’s decision‑making process within its documents and governance.

“You can call that an agent company or AI-enabled governance,” he said.

While researching, he found numerous efforts around digital identity and AI agents, including work by major tech firms, but could not find a standard that linked both an AI agent and a company together as a single credential.

“In tech, when you develop new versions of things, there is usually a standard,” he said. “If you are going to have 10,000 developers all producing cryptocurrency or an agent, typically they will agree to a format for the data so when they interact, they are interoperable.”

He turned to Anthropic’s large language model Claude to create a standard. Within ten minutes Claude wrote a white paper for him. He released the first version of the paper a week ago seeking feedback. After gaining that, he put out a second version with refinements.

So far one firm and two individuals have expressed interest in using the standards.

“One person who reached out was an Australian fintech entrepreneur, quite well known,” Mr Collier said. “He was in an expert panel for the Obama administration.”

He is now looking for people to join a technical working group to guide the next version of the paper. The group would be a good fit for Bermuda financial sector stakeholders, particularly insurers, investment funds and others would use a standard like this.

For now, the standard remains an early‑stage proposal.

With AI agents rapidly moving into finance and commerce, Mr Collier said the need for clear ways to identify and bind them to accountable humans and companies will only grow.

“Once agents start moving real money around, regulators and lawyers will pay very close attention,” he said. “We are going to need standards like this sooner rather than later.”

•To contact Mr Collier e-mail info@aiagentservices.net or contact him through GitHub or LinkedIn