Ten artykuł nie został jeszcze przetłumaczony na Twój język. Wyświetlamy go po angielsku. Przeczytaj oryginał po angielsku

Regulacje2 min czytania0 komentarzy

HKMA studies AI agents to trace illicit stablecoin flows

Raymond Chan said a single on-chain agent could combine analytics, due diligence and freezing, citing $154B in illicit receipts for 2025.

Regulacje
Opublikowano
HKMA · STABLECOINS: $154B

Najważniejsze

  • The HKMA is studying agentic AI to help banks monitor digital-asset flows for money laundering.
  • Raymond Chan said stablecoins are a preferred vehicle for illicit activity because of stability and self-custody transfers.
  • Chan cited Chainalysis data: over $150 billion reached illicit addresses in 2025, 84% in stablecoins.
  • The report's original figure is at least $154 billion, up 162% from a revised $57.2 billion in 2024.
  • The HKMA's SAGE project, launched in November, is setting identity and governance rules for AI agents.

The HKMA's case for AI agents

The Hong Kong Monetary Authority is studying how banks could use agentic artificial intelligence — AI systems that carry out tasks on their own — to monitor digital-asset activity. Raymond Chan, the authority's Executive Director for Enforcement and AML, set out the idea at a seminar the HKMA hosted on 17 November. It was the third event in a series on bringing AI into financial-crime monitoring; this one focused on digital assets.

Chan described digital assets as combining speed, global reach, pseudonymity and limited reversibility. Once funds reach a fraud group or leave an exchange after a hack, they can cross borders within minutes and bypass traditional intermediaries, which makes recovery and interception harder. He singled out stablecoins: their price stability and peer-to-peer transfers through self-custody wallets make them a preferred vehicle for illicit activity, he said.

Five tools in one agent

Chan sketched what a single on-chain AI agent could combine: blockchain analytics, customer due diligence records, fraud intelligence, suspicious-wallet monitoring and rapid freezing. The hard part, he said, is linking wallet activity to a real-world identity fast enough, because funds disperse quickly when that link is slow.

He said no single tool provides complete protection and that defences have to be layered. On stablecoins, the HKMA is taking a cautious approach that puts investor protection and financial-crime risk management first, alongside a broad monitoring strategy. Technologies aimed at pseudonymity, such as decentralised identity, are at the exploration and pilot stage, according to Chan.

The figures behind the plan

Chan cited Chainalysis data showing that illicit addresses received more than $150 billion in crypto in 2025, with stablecoins making up 84% of that total. The Chainalysis Crypto Crime Report 2026 puts the original figure at at least $154 billion, up 162% from a revised $57.2 billion in 2024. The report attributes most of the increase to receipts by sanctioned entities, which rose 694%.

The same report says illicit activity still accounts for less than 1% of attributed on-chain transaction volume. The HKMA's SAGE project, launched in November, is setting identity-verification and governance rules for AI agents before they are deployed.

Źródła

2 źródła
  1. Crowdfund InsiderHong Kong Monetary Authority (HKMA) Highlights Agentic AI Potential for Strengthening Anti-Money Laundering Controls on Digital Assets
  2. BlockTempo 動區動趨香港金管局想養 AI 代理抓洗錢!穩定幣佔非法資金 84%

Ten artykuł zredagował Clause, reporter AI BattleTown News, na podstawie raportów wymienionych powyżej.

Najczęstsze pytania

What is the HKMA proposing agentic AI would do?

The HKMA's Executive Director for Enforcement and AML, Raymond Chan, described agentic AI as a way to combine blockchain analytics, customer due diligence data, fraud intelligence, suspicious-wallet monitoring and rapid freezing in one on-chain agent. He said no single tool gives full protection and that controls must be layered.

How large are illicit crypto flows, according to the figures cited?

Chan cited Chainalysis data that illicit addresses received more than $150 billion in crypto in 2025, with stablecoins accounting for 84%. The report's original figure is at least $154 billion, up 162% from a revised $57.2 billion in 2024.

Ten artykuł ma charakter wyłącznie informacyjny. Nie jest poradą finansową, inwestycyjną ani prawną. Zastrzeżenie

Komentarze: 0 komentarzy

BattleTown: bitwy czołgów w przeglądarce

Nowe minigry, „Szturm” i „Przełom”, już w BattleTown

Wróć do listy

Czytaj także

Wpisz co najmniej 2 znaki