0x1c47a0dd1c47…1c47a0e0
Hong Kong's HKMA to Study Agentic AI for Digital-Asset Tracking
The HKMA will study agentic AI to trace digital-asset transactions and plans a layered system to help banks match on-chain transfers with off-chain customer data and flag abnormal fiat-crypto flows.

Outputs
The HKMA will study agentic AI to track digital-asset transactions.
Assistant Chief Executive Chan King-hung, covering regulation and AML, announced the study.
A separate multilayer system will let banks compare on-chain transfers with off-chain customer data.
Chan said permanent public blockchain records can help regulators and law enforcement freeze suspicious funds quickly.
No deployment timeline has been published for either initiative.
The Hong Kong Monetary Authority (HKMA) will study agentic artificial intelligence to track digital-asset transactions, and separately plans a multilayer monitoring system to help banks identify unusual flows between digital assets and fiat currency.
HKMA Assistant Chief Executive Chan King-hung, whose portfolio covers regulation and anti-money laundering, announced the AI study. The initiative places Hong Kong's central banking institution among a growing set of monetary authorities examining whether autonomous, goal-directed AI agents can perform supervisory tasks that currently demand manual investigation.
What does the technology change for supervisors?
Chan's reasoning rests on a structural property of public blockchains: permanence. He said blockchain transfers leave permanent public records. Combined with industry customer due diligence and data used to trace false information, those records can give regulators and law enforcement leads to suspicious funds and help them freeze the funds quickly.
That framing matters for practitioners. The supervisory bottleneck in digital-asset AML has never been the absence of data — distributed ledgers record every transfer in the clear — but the cost of converting raw transaction trails into actionable leads before funds move on. Chan's proposal treats agentic AI as the conversion layer: a system that reads immutable on-chain records against the customer files banks already maintain and surfaces what a human analyst would otherwise assemble by hand.
The emphasis on rapid freezing signals where the authority sees the operational payoff. Suspicious-fund investigations fail most often on speed, not detection; the window between identifying tainted flows and immobilizing them is measured in hours.
How would the layered tracking system work?
In a separate measure, the HKMA plans a multilayer tracking system to monitor on-chain transactions. The design has a concrete institutional purpose: supporting banks in comparing blockchain transfers with off-chain customer transaction data.
For compliance teams, that comparison is the hard part of digital-asset AML. A bank sees its own customer's fiat activity internally, but the corresponding on-chain movement sits on a public network the bank does not control. The planned system would bridge the two views, spotting abnormal flows between digital assets and fiat currency early — before suspicious activity matures into a full-scale laundering case.
Early detection also changes the economics of compliance. Flagging anomalous fiat-to-crypto and crypto-to-fiat flows at inception reduces the volume of retrospective reviews, alert backlogs and late-stage suspicious transaction reports that currently dominate bank AML budgets.
What is the regulatory balance?
The authority framed its regulatory approach as balancing innovation in digital finance with anti-money-laundering requirements. That positioning is deliberate. Hong Kong has courted the digital-asset sector with a licensing regime for virtual-asset platforms and tokenization initiatives, while its banks remain the chokepoint through which digital-asset businesses access fiat rails.
For licensed institutions and digital-asset firms operating in Hong Kong, the message is that surveillance capability will scale alongside market access. The same permanent public records that Chan cited as an investigative asset are also the substrate banks will be expected to reconcile against customer files under the layered system.
The AI study and the tracking system remain at the planning stage, with no published deployment date. Banks and virtual-asset service providers in the jurisdiction should expect supervision tools that arrive before, not after, the next wave of digital-finance products.
via tokenpost.com (Source)
More from Nathan Brooks
Show full bio
Market editor covering money, funding and market shifts around crypto and web3 industry at Mempool Brief.
1,512 articles