The Synthetic Supplier: How AI Is Industrialising International Trade Fraud

An Irish food distributor receives an enquiry from a hotel-supply company in the Middle East. The prospective customer wants to purchase three refrigerated containers of premium dairy products for a new hospitality development.

The opportunity is worth €420,000.

Over the next four weeks, everything appears credible.

The buyer has a professional website, an active LinkedIn presence and photographs of completed hotel projects. Its purchasing director joins two video calls and speaks confidently about product specifications, import requirements and shipping schedules.

The exporter receives a certificate of incorporation, trade licence, customer references, bank letter and signed purchase order. The buyer negotiates firmly but reasonably, pays a deposit and requests 60-day credit for the balance.

Shortly before collection, the exporter is told that the buyer has appointed a different freight forwarder. Revised instructions arrive from what appears to be a legitimate logistics company.

The warehouse releases the containers.

Days later, the freight forwarder denies involvement. The deposit is traced to a compromised third-party account. The corporate documents are sophisticated fabrications. The LinkedIn profiles are built around synthetic identities. The executive on the video calls was not who he claimed to be.

The goods are gone.

This example is fictional, but every element of it reflects capabilities already available to organised fraudsters.

Artificial intelligence is not inventing international trade fraud.

It is making it faster, cheaper, more convincing and easier to scale.

Welcome to the age of the synthetic supplier.

Trade fraud has moved beyond the fake invoice

Most businesses know to question an unexpected request to change bank details or an invoice sent from a slightly unfamiliar email address.

That model of fraud is now dangerously incomplete.

Generative AI can help criminals create an entire ecosystem of mutually supporting evidence:

• Professional websites and company histories

• Synthetic directors and employee photographs

• Personalised emails in multiple languages

• Product catalogues and factory imagery

• Altered financial records and payment confirmations

• Fabricated licences and inspection reports

• Cloned voices and manipulated video conversations

The result is not one forged document. It is a coherent commercial fiction.

INTERPOL’s 2026 Global Financial Fraud Threat Assessment estimates that AI-enhanced fraud can be 4.5 times more profitable than fraud conducted without AI. It also warns that emerging agentic AI systems could help criminals plan and execute entire fraud campaigns.

Europol’s 2026 Internet Organised Crime Threat Assessment similarly reports that generative AI is enabling criminals to personalise social-engineering attacks, accelerate fraud and operate across borders more effectively.

This is the industrialisation of deception.

When every piece of evidence agrees

International trade depends heavily on documentary trust.

A buyer may never visit the supplier’s factory. An exporter may never meet its overseas distributor. A freight forwarder may collect valuable goods based on electronic instructions from people it has never encountered.

Confidence is therefore built through proxies.

A registration certificate suggests that the company exists. A bank letter suggests financial credibility. An inspection report suggests the goods meet specification. A video call suggests that the person on the screen is real.

But what happens when AI can manufacture every proxy?

The greatest danger is not that one document looks suspicious. It is that every fraudulent element appears to corroborate everything else.

The website supports the company profile. The salesperson supports the website. The LinkedIn identity supports the salesperson. The inspection report supports the product photographs.

Everything looks consistent because everything was created as part of the same deception.

Convincing documents are becoming harder to detect

Businesses often assume that fraudulent documents will contain obvious warning signs: poor formatting, mismatched fonts, distorted logos or visible editing.

That assumption is becoming unsafe.

The 2026 AIForge-Doc research benchmark tested detection systems against 4,061 AI-forged financial and form documents across nine languages. Existing document-forensics tools deteriorated substantially when faced with AI-generated alterations. One multimodal AI evaluator performed little better than chance.

A separate study, GPT4o-Receipt, examined 1,235 authentic and AI-generated receipts. Human reviewers often struggled to identify the synthetic documents because some of the strongest clues were not visual defects, but arithmetic inconsistencies hidden inside otherwise convincing records.

Staring more closely at the PDF is no longer a fraud-prevention strategy.

A genuine transaction can still be hijacked

The supplier does not even need to be fake.

Criminals can impersonate freight forwarders, carriers, customs agents or finance employees. They can insert themselves into genuine email chains, imitate legitimate domains and issue fraudulent payment or collection instructions.

The TT Club and BSI Consulting 2024 Cargo Theft Report found that strategically planned theft represented 18% of reported US cargo-theft incidents. These cases involved tactics such as carrier impersonation, false documentation and identity fraud.

The supplier may be genuine while the carrier is fake.

The carrier may be genuine while the collection instruction is fraudulent.

The transaction may be legitimate while the email changing the beneficiary bank account is not.

The FBI’s 2024 Internet Crime Report recorded more than 859,000 complaints and losses exceeding $16 billion, a 33% increase over the previous year. Business email compromise remained one of the most financially damaging categories.

International trade is particularly attractive to criminals because it combines high-value payments, unfamiliar counterparties, complex documentation, multiple handoffs and constant time pressure.

SMEs are especially exposed

Small and medium-sized businesses are encouraged to diversify suppliers and enter new markets, yet many operate without dedicated fraud, compliance or procurement-security teams.

Due diligence may consist of a website review, company search, emailed certificates and references supplied by the counterparty. Important negotiations may take place through personal inboxes or WhatsApp.

At the same time, legitimate SMEs continue to struggle to obtain finance. The Asian Development Bank’s 2025 Global Trade Finance Gap Survey estimated unmet global trade-finance demand at $2.5 trillion and reported that 41% of SME applications were rejected.

This creates a dangerous environment. Genuine businesses are under pressure to find new financiers, suppliers and intermediaries, while fraudsters can manufacture increasingly convincing alternatives.

Digital does not automatically mean trustworthy

Uploading a fraudulent document into a digital repository does not make it authentic.

Scanning a forged certificate does not validate it. Digitising an invoice does not prove that the issuer was authorised. Moving a conversation from email into software does not verify the identity of the participants.

Digital trade is not necessarily trusted trade.

Trust requires evidence of identity, authority, provenance and accountability across the entire transaction.

Businesses need to know who supplied each piece of information, whether participants were independently verified, whether documents changed, who approved each milestone and whether payment instructions belong to the contracting entity.

Trust must become a system

The answer is not simply more paperwork.

Every additional certificate creates another object that can be forged.

The stronger response is a structured transaction environment in which participants, documents, approvals, obligations and changes are connected.

This is where platforms such as iTradeDigital can play an important role: creating a single source of truth, linking documentation to transaction milestones, enabling controlled collaboration and preserving a traceable record of who did what and when.

Technology cannot eliminate fraud. But it can make contradictions harder to conceal and unauthorised changes harder to introduce.

For generations, international trade has depended on documents as proxies for trust.

AI can now manufacture those proxies on demand.

In the age of the synthetic supplier, trust can no longer be assumed because the paperwork looks convincing.

It must be designed into the transaction itself.