Sharplink CEO Joseph Chalom said no Agent, platform, or financial institution should become the unavoidable gateway for people entering financial markets. In his view, the objective is not just to create more AI Agents. It is to build an interoperable financial system in which Agents compete with one another and users can move between platforms freely, without control being concentrated at a small number of key access points.
This is the third article in his series on Agent finance. The first looked at the rise of a new Agent economy and its implications for finance. The second focused on the coming battle over financial service fees and how consumers could benefit. This latest piece asks whether that shift can avoid a familiar outcome from earlier technology cycles: a market eventually dominated by a handful of winners.
Chalom argues safeguards need to be built before Agent finance scales
Chalom wrote that much of today’s AI debate centers on open versus closed models, whether governments should slow the pace of AI development, and competition between China and the United States for AI leadership. But he said those debates do not answer a more immediate question: how people’s financial security will be protected as Agent finance becomes more common over the next few years.
He added that the world’s largest financial and technology companies have already started positioning for that competition. With so much attention fixed on broader questions about AI, the financial safeguards that need to be put in place now could be overlooked.
His view on regulation was not purely dismissive. Chalom acknowledged that business leaders often react negatively to regulation, especially in emerging industries, and that regulators sometimes move too quickly before they fully understand a technology. Even so, he argued that carefully designed and properly implemented rules can build trust, set standards, and support growth.
From there, he laid out a set of principles that he said could create the trust needed for Agent economy development to proceed responsibly.
Principle one: Agents must serve users
Chalom said the obligations an Agent owes to a user need to be made explicit. An Agent should act in the interest of the individual or business that entrusted it with a task. Its recommendations should match the customer’s needs, not the interests of outside third parties.
He also said all commercial relationships must be transparent. Customers should be able to understand how an Agent and the company behind it make money. Any proposed contract or partnership arrangement, he wrote, should fully disclose relevant financial information and conflicts of interest.
He drew another clear line around delegated work. Handing a task to another Agent, he said, must never become a way to get around the original Agent’s limits or the user’s existing preferences.
Principle two: users must remain in control at all times
Authorization, in Chalom’s framework, has to be explicit and bounded. Customers should decide what an Agent can and cannot do, which activities are permitted, how long approval lasts, which accounts may be accessed, and what spending limits apply.
Users also need the ability to revoke permission. Every Agent, he wrote, should have a mechanism that can stop its actions immediately, and that mechanism must also extend to any authority the Agent has delegated onward. Once permission is withdrawn, the Agent must comply at once.
Every action should also have a clearly identifiable responsible party. Chalom said actions taken by an Agent and by any delegated Agent must leave an auditable record, and each decision should be traceable to the individual or company that created or authorized that Agent.
Principle three: Agents must be portable
Chalom argued that Agents should be free to use services from different providers. Both users and Agents, he wrote, should be able to connect to competing providers and switch between them.
That portability should also apply to data. Agents should be able to carry financial records, preferences, settings, and identity credentials to a new provider. Users, for their part, should be able to delete or remove personal data in line with applicable law.
He extended the same logic to identity. Agents should be able to keep a persistent identity that works across platforms. Chalom described that identity as a unique non-human digital identifier paired with an access-permission configuration, allowing an Agent to preserve continuity of identity across tasks, retain memory over time, and keep accountability auditable.
He says action is needed now, not later
Chalom called on responsible industry participants and policymakers to put safeguards and trust mechanisms in place now so the Agent sector can develop in an open and fair environment. He also said those principles should become part of the wider AI policy discussions already under way in governments around the world.
He did not present that as easy work. Technology, he wrote, always moves faster than rulemaking. Regulating AI may be the hardest policy problem lawmakers have faced in decades, and the pace of AI innovation has already outstripped earlier technology waves.
Still, he warned that failing to act now would leave incumbents and large technology companies to write the rules themselves, deciding how they interact with Agents and consumers. Based on past experience, he said, those rules could end up embedded in binding terms of service, click-through agreements, and other hard-to-see competitive barriers.
Ethereum is cited as one possible neutral foundation
Chalom closed by linking those policy principles to the infrastructure layer. Even with the best possible regulatory framework, he wrote, it is difficult to imagine an Agent economy meeting those standards without open and decentralized systems underneath it.
He specifically pointed to neutral blockchains such as Ethereum as among the strongest candidates to provide that environment. In that setting, developers, entrepreneurs, and users would be free to build new Agent functions and launch new businesses. He also said such blockchains are among the most credible paths to expanding the potential market for Agent finance and passing benefits on to consumers.
At the end of the article, Chalom described himself as an optimist who believes Agent finance can improve the lives of billions of people around the world. He also called himself a realist, adding that the shift has to be shaped deliberately from now on if it is to produce a better outcome than earlier technology cycles.