Interpol: AI scams earn fraudsters 4.5 times more than traditional scams

Fraudsters using artificial intelligence tools earn significantly more than those working through traditional methods. According to Interpol, such operations can generate 4.5 times more than classic fraud, and this is rapidly changing the very economics of digital deception.

Why income is higher

The key difference of AI schemes isn't that they look "futuristic," but that they dramatically reduce costs for criminals. What previously required hours of manual work—writing emails, selecting wording, imitating business tone, creating fake profiles and communication scenarios—can now be done in minutes. As a result, a fraudster spends less time on one attack and simultaneously launches more attempts without increasing their team or budget.

The second reason is the quality of contact with the victim. Generative models help create more persuasive, personalized, and well-written messages. If bulk spam used to give itself away through poor language, strange logic, or templated text, now text, voice, or images can look far more believable. This increases conversion: more people respond, click links, share data, or agree to transfer money. Fewer mistakes, more victim trust.

How schemes are changing

AI doesn't necessarily invent new types of crimes—it amplifies already familiar schemes and makes them scalable. Essentially, criminals get an inexpensive tool that helps them test different approaches faster, adapt to victim reactions, and mass-produce convincing content for a specific scenario. What previously required a team of copywriters, call center operators, or lengthy manual preparation can now be assembled almost on a conveyor belt and launched without pauses.

• Phishing emails and messages without obvious language errors
• Fake messages impersonating a bank, colleague, or relative
• Voice and video deepfakes for urgent requests to transfer money
• Mass creation of fake profiles and support services
• Quick localization of schemes to another language, region, or audience

Because of this, the barrier to entry for fraudsters is lowered. To assemble a plausible attack, you no longer need to know the language well, write convincing texts, or spend a long time preparing a cover story. It's enough to correctly formulate a prompt to the model, upload public information about the victim, and automate distribution. For organizers of such schemes, this means higher margins; for regular users, it means fewer visual and linguistic cues that used to make it easy to quickly spot deception.

Where the main risk is

The danger isn't only that AI makes old schemes more effective, but the speed of their adaptation. If a certain cover story stops working, it can be quickly rewritten, tone changed, language and story details altered, and the campaign launched again. This hypothesis-testing cycle is especially profitable in fraud: criminals learn from people's reactions almost in real time and improve scenarios almost automatically with minimal costs.

For business, this is a separate problem. Employees increasingly receive messages that externally resemble ordinary work requests: a request to urgently pay an invoice, change payment details, open a document, confirm a code, or call a manager. When such messages are written without errors, consider company context, and even come with synthesized voice, human caution alone no longer provides the previous level of protection. This requires verification processes, not just reliance on attentiveness. This is why the focus shifts from recognizing "suspicious text" to verifying identity and intent.

If a request involves money, access, confirmation codes, or urgent actions, verification should go through a separate communication channel. The more convincing AI tools become for attackers, the more important simple rules become: double-check, don't rush, and don't trust a single message just because it sounds professional. Many attacks now count on exactly this.

What this means

Interpol's assessment shows something straightforward: AI has become not just a productivity tool, but an amplifier of criminal economics. While defensive practices at companies and by users change more slowly, fraudsters gain an advantage. Therefore, basic digital hygiene, multi-factor authentication, internal regulations, and employee training turn from a useful recommendation into a mandatory minimum. Without this, any convincing fake, urgent call, or plausible email will work significantly more often.