A United Nations report on organized online fraud treats artificial intelligence as one of several technologies that scam operators have adopted, and it does not measure how much AI adds to losses. The report, titled "Inflection Point" and published in April 2025 by the UN Office on Drugs and Crime (UNODC), says scam operations use large language models for translation and deepfake filters for video calls and advertisements. It also describes cryptocurrency as one of the channels through which the proceeds move.
Key takeaways
- UNODC estimated that cyber-enabled fraud caused financial losses of USD 18 billion to USD 37 billion in East and Southeast Asia in 2023.
- AI appears in the report as translation, deepfake filters, and software sold on illicit online marketplaces, and the report gives no figure for AI's share of scams.
- The report cites US authorities reporting more than USD 5.6 billion in cryptocurrency scam losses in 2023.
- Its recommendations stress laws, oversight, and enforcement capacity for money laundering, virtual assets, and casinos, not limits on AI.
- This article is general information and not investment advice.
The report and its scope
The UNODC technical policy brief comes from the agency's regional office for Southeast Asia and the Pacific. It covers scam centers, underground banking, and illicit online marketplaces, and it draws on data shared by governments, international partners, and other organizations. The United Kingdom government funded the research. The focus is organized crime in the Mekong region and its spread elsewhere, so the findings describe one set of groups and are not a census of all online fraud.
Where AI appears in the findings
The executive summary says that crime groups in the region have integrated technologies including malware and artificial intelligence into their operations. A later section adds that groups responded to law enforcement pressure by using blockchain, cloud computing, generative AI, and machine learning, which gave them new fraud capabilities. Specific uses named in the text are large language models for translation and deepfake filters for video calls and advertisements.
The report also describes illicit online marketplaces that sell services to other criminals. Their listings include stolen data lists, compromised accounts, pre-registered SIM cards, fraud scripts and videos, and AI tools such as automation and deepfake software. Deepfake, sextortion, and malware-enabled fraud incidents have increased sharply across East and Southeast Asia in recent years, according to the report, with growing attribution to local criminal networks.
The text keeps the human side in view as well. It says that workers who speak the target's language and understand cultural quirks make a pitch more convincing, so technology supplements staff rather than replacing them.
Where crypto-assets appear
Virtual assets run through the report as part of money movement. It links them to casinos, online gambling, and underground banking, and it recommends reviewing laws on money laundering and virtual assets. In one case it describes, authorities in Viet Nam reported in August 2024 that they had disrupted a cell of a regional laundering organization that handled unauthorized peer-to-peer cryptocurrency services totaling more than 1.1 billion units of a dollar-linked stablecoin in transactions between November 2023 and May 2024. Bank accounts and front company registrations were part of the same operation. In the cases the report describes, virtual assets sit alongside cash, bank accounts, front companies, and property.
What the report says and what it leaves open
| Topic | What the report says | What it does not say |
|---|---|---|
| AI use | Translation by language models, deepfake filters, and AI tools sold on marketplaces | No measure of how many scams use AI or how much AI changes success rates |
| Losses | USD 18 billion to USD 37 billion in East and Southeast Asia in 2023 | No split by technology used |
| Crypto | Virtual assets used with banks, casinos, and front companies to move proceeds | No claim that cryptocurrency causes the fraud |
| Response | Review laws on money laundering and virtual assets, supervise high-risk financial providers, build enforcement skills | No call to restrict AI tools |
Reading headline numbers with care
The figures are estimates, they cover 2023, and they describe a region. The report also cites US authorities reporting more than USD 5.6 billion in cryptocurrency scam losses in 2023, with an estimated USD 4.4 billion attributed to so-called pig butchering schemes, which it says are most prevalent in Southeast Asia. A statement about a surge in AI-driven scams is more reliable when it answers these questions:
- Which year and region does the number cover?
- Is the loss figure reported by victims, estimated by an agency, or both?
- Does the source measure the share of cases that involve AI, or only describe AI as present?
- Does a claim about cryptocurrency separate where money was paid from where it was laundered?
This article summarizes a public report. It is not investment, financial, or legal advice, and it does not recommend or assess any token, exchange, wallet, or service. A person who thinks they have been targeted by a scam can contact their local police or a national fraud reporting service.
FAQ
Does the report say AI causes most scam losses?
No. It names AI among several technologies, along with malware and blockchain, and it gives no estimate of AI's contribution to losses.
Does the report blame cryptocurrency for these scams?
The report describes organized-crime operations in which virtual assets are one of the channels for moving money. It recommends reviewing laws on money laundering and virtual assets next to oversight of casinos and special economic zones.
Are there newer figures?
Later UNODC work may update the estimates. This article relies on the April 2025 report, which was checked at the source, and the figures should be read as 2023 estimates.
Related reading
- How AI and cryptocurrency differ
- Risks and pitfalls of cryptocurrency
- Documented concerns about Bitcoin
- Where AI helps, with evidence
More in AI in finance.
Last verified: October 2026
Sources
- Inflection Point: Global Implications of Scam Centers, Underground Banking and Illicit Online Marketplaces in Southeast Asia (technical policy brief), United Nations Office on Drugs and Crime, April 2025