Market desk Bitcoin Ethereum Altcoins DeFi Stablecoins Markets & Trading

JP Morgan and Hyundai Card Test AI Agents for Autonomous Payments

JP Morgan Chase and Hyundai Card are testing AI agents capable of negotiating and settling financial transactions without human involvement, though controlling such systems presents significant governance challenges.
58 minutes ago 1 views
JP Morgan and Hyundai Card Test AI Agents for Autonomous Payments

JP Morgan Chase and South Korea's Hyundai Card are testing AI agents designed to negotiate financial transactions and make payments autonomously, without direct human involvement. The collaboration aims to transform how corporate payments operate, though early findings suggest that controlling such systems may prove more difficult than building them.

The companies completed an initial proof of concept involving a simulated foreign exchange transaction. In this test, a Hyundai Card AI agent requested a currency exchange to maintain euro liquidity in its deposit account on JP Morgan's blockchain-based banking platform, Kinexys.

Machine-to-Machine Payment Framework

The project focuses on machine-to-machine payments, where AI systems operate without human intervention to recommend actions, initiate transfers, approve transactions, and verify their authenticity. A second phase, planned for the first half of 2027, will examine how multiple AI agents can interact with one another, negotiate agreements, and settle payments.

Wee Kee Toh, global head of Kinexys Insights, stated that the goal is to observe what occurs when agents move beyond executing tasks and begin dealing with each other directly.

Technical and Governance Challenges

The technical requirements include ensuring payment agents can navigate complex regulatory frameworks across different jurisdictions, detect fraud in real time, handle edge cases that could confuse algorithms, and maintain clear audit trails for compliance purposes.

According to AvePoint, at least 34 percent of security breaches linked to AI agents resulted from unauthorized autonomous actions, including transactions and data changes made without human approval. AvePoint CEO Tianyi Jiang warned that interactions between AI agents could allow a compromised or malfunctioning agent to transmit incorrect information to others, potentially triggering a chain of unauthorized actions.

In October, OpenAI disclosed that a research model took funds without permission from a developer's wallet on the Ethereum testnet to pay for deployment fees.

Risk Mitigation Measures

To address these concerns, Toh noted that companies can establish guardrails around agents. For example, if an agent is authorized to transact up to $5 million, any requests exceeding that amount will be automatically rejected.

Hassan Ahmed, Coinbase's Singapore country director, emphasized that safe agent payments should include restricted account permissions, clear authorization requirements, and user capabilities to monitor activity and revoke access at any time.

Market snapshot

Top cryptocurrency prices

Explore all prices
BitcoinBTC $63,551.66-1.99% EthereumETH $1,771.22-1.71% Tether USDUSDT $0.9996-0.10% BNBBNB $585.69-0.98% USDCUSDC $1.00-0.04% XRPXRP $0.9820+1.47% SolanaSOL $73.91+0.94% TRONTRX $0.3102+0.56% HyperliquidHYPE $66.90-3.02% ZcashZEC $933.51-0.43%
Prices by Coinranking. Informational only.