JPMorgan, Hyundai Card test autonomous AI transactions on blockchain

JPMorgan, Hyundai Card test autonomous AI transactions on blockchain

JPMorgan Chase and South Korea’s Hyundai Card have completed a proof-of-concept trial in which artificial intelligence agents executed a simulated foreign exchange transaction without direct human intervention. The initiative, which utilizes JPMorgan’s Kinexys blockchain platform, represents a step toward autonomous financial operations where AI agents may eventually negotiate and settle payments with one another.

The initial test involved a Hyundai Card AI agent requesting a foreign exchange transaction to maintain euro liquidity within a Kinexys blockchain deposit account. JPMorgan described this milestone as the first private blockchain-based autonomous bank-agent-to-client-agent test for corporate treasury management. Wee Kee Toh, global head of Kinexys Insights at Kinexys by JPMorgan, stated that the next phase aims to expand testing to more complex scenarios involving multiple interacting agents.

The companies are targeting the first half of 2027 for this expanded testing period. Toh noted that the challenge is less about the technical feasibility of the AI and more about establishing robust governance structures and transaction limits. Kinexys has developed a framework that restricts AI agents to predefined parameters, automatically rejecting any requests that fall outside authorized limits to mitigate risks associated with autonomous decision-making.

10 responses to “JPMorgan, Hyundai Card test autonomous AI transactions on blockchain”

  1. The 2027 timeline feels incredibly optimistic. Multi-agent systems are still hitting major coordination headaches.

  2. Why use a private blockchain when public ones already handle this? JPMorgan is just building a walled garden.

  3. Autonomous agents settling trades is scary cool. Just hope the governance limits actually hold up under stress.

Leave a Reply

Your email address will not be published. Required fields are marked *