No Human Intervention Needed! JPM Teams Up With Hyundai Card to Explore AI Agents in Autonomous Financial Transactions
JPMorgan and Hyundai Card of South Korea are testing how AI agents handle financial transactions without direct human intervention, including whether they can negotiate with each other and complete payments in the future.
According to Zhitong Finance APP, artificial intelligence (AI) agents have already demonstrated that they can play a positive role, create significant trouble, or sometimes do both. One of the areas believed to have immense potential is payments. If enough trust can be established, AI agents could manage a range of financial matters on behalf of individuals. For this reason, JPMorgan (JPM.US) and South Korea's Hyundai Card are testing how AI agents could handle financial transactions without direct human intervention, including whether they could negotiate with each other and complete payments in the future.
The two companies have already completed a preliminary proof of concept involving a simulated foreign exchange transaction and plan to launch a more complex testing phase in the first half of 2027, which will involve multiple AI agents.
Wee Kee Toh, Global Head of Kinexys Insights, a subsidiary of JPMorgan focused on blockchain, stated that the next step is to test what happens when autonomous agents interact with each other within an ecosystem, beyond just executing transactions. He said, "What happens when agents start negotiating with each other, reaching agreements directly, and then making payments? What does that process actually look like?" "I believe this is actually the most crucial part we want to expand upon based on the scope of the first phase of testing."
In the initial test, the Hyundai Card AI agent initiated a foreign exchange request to maintain euro liquidity in its Kinexys blockchain deposit account. JPMorgan stated that this was the first enterprise cash management test on a private blockchain, with autonomous bank and client AI agents interacting.
Accountability Issues
Although the initial test has been completed, JPMorgan has not yet set a timetable for the commercialization of this technology.
Wee Kee Toh stated that the technical build itself is relatively straightforward, but establishing appropriate governance and internal controls is much more challenging. He said, "From a technical perspective, this is actually not the biggest hurdle." He pointed out that the real issue to address is accountability, and the risks of allowing AI agents to execute financial transactions without direct human oversight.
Experts say that such agents have the potential to change how the financial services industry interacts, but this transformation must come with corresponding accountability mechanisms. In its "AI Status Report 2026," data management and protection software company AvePoint found that unauthorized autonomous agent behavior is the third most common AI-related security vulnerability, accounting for 34.1% of enterprises reporting such incidents. This behavior includes executing trades or modifying data without approval, though the survey's findings are not limited to corporate cash management payments.
AvePoint CEO Tianyi Jiang commented, "When agents are negotiating and operating autonomously, the focus of governance shifts from monitoring human actions to monitoring dependencies between machines." He explained that interactions between AI agents could amplify errors and trigger transactions before human intervention is possible.
He said, "The worst-case scenario is a cascading failure. An agent that is compromised or hallucinating could pass erroneous data to another agent." He added that this could lead to "a series of incorrect financial transfers executed rapidly, even before humans are aware of the problem."
To address these concerns, Kinexys has developed a framework that confines AI agents' activities within preset transactional parameters and automatically rejects requests beyond the authorized limits. Wee Kee Toh stated, "You need to draw a boundary for the agent. The agent can never do anything beyond that boundary." For example, if an AI agent is authorized to execute transactions up to $5 million, any transaction exceeding this threshold would be blocked by the system.
Hassan Ahmed, Singapore Country Director at Coinbase (COIN.US), noted that in order to ensure the security of agent payments, clear safeguards must be in place, including spending caps, account access limits, and clear authorization requirements. He said that, given AI may create new opportunities for malicious actors, users should also be able to monitor activity in real time and revoke access at any time.
Hassan Ahmed stated, "While in certain use cases, private or permissioned environments may offer additional controls, we must not overlook the potential of open blockchain networks to support a more inclusive, interoperable, and efficient financial system."
Beyond the Pilot Stage
Wee Kee Toh mentioned that JPMorgan and Hyundai Card are also exploring the possibility of expanding their collaboration to other businesses within Hyundai Motor Group. The broad business scope of Hyundai Motor Group is one reason JPMorgan chose Hyundai Card as a partner. The two sides have not yet announced any specific projects.
According to Hyundai Card's website, Hyundai Motor Group, through Hyundai Motor, Kia, and Hyundai Commercial, holds a combined 78.04% stake in Hyundai Card.
Previously, Hyundai Card has tested blockchain-based cross-border payments. In July of this year, the company said it completed a proof-of-concept test, transferring $20,000 via USDT stablecoin between Hyundai Motor's US and Mexico subsidiaries, with the entire transfer and verification process averaging seven minutes.
Kinexys is also continuously expanding its blockchain-based payments business. In August this year, the JPMorgan business unit announced that the cumulative value of transactions processed since its inception had exceeded $4 trillion, with a daily average of about $7 billion. In addition, in June 2025 JPMorgan launched a proof of concept for its USD-denominated deposit token JPMD and began offering the product to institutional clients in November of the same year.
Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.
You may also like
US Dollar Index (DXY) takes a breather after rally to 19-month high

BTCS warns of forced ETH sales as Ethereum tests $1,800 support
World Bank to Explore USDC Stablecoin for International Payments

