Interpol: AI scammers earn 4.5 times more and reach a new scale

The use of artificial intelligence is already noticeably changing the economics of online fraud. Schemes in which criminals leverage AI generate on average 4.5 times more profit, and Interpol warns that as AI agents develop, such operations could scale to near-industrial levels.

Why Income is Growing

The main effect of AI for fraudsters is not magic but sharp cost reduction and acceleration of familiar tactics. Where previously manual work was needed to write dozens of email variations, correspond with victims, and adapt messages for language, tone, and context, now a model handles a significant portion of that work. It generates text faster, adjusts style for specific individuals, and enables more attempts in the same time frame. As a result, old schemes become more efficient without proportional increases in spending on personnel and infrastructure.

The 4.5x profit increase speaks not only to higher-quality deception but also to a new economy of scale. Where fraudulent campaigns previously hit ceilings imposed by operator availability, preparation time, and language barriers, AI removes multiple constraints at once. Particularly dangerous is how automation enables rapid testing of phrasings, scenario changes, and reuse of successful templates. For criminals this means broader reach, higher conversion rates, and less manual routine work.

What AI Agents Change

The next risk extends beyond simple text or voice generation to AI agents capable of executing action sequences almost autonomously. Such a tool can not only write a message but also gather data, send messages, conduct dialogue from a script, and adapt as the conversation unfolds. If these capabilities migrate to criminal scenarios, fraud will cease to be a collection of isolated episodes and will become a managed process with minimal human involvement at each step.

Interpol fears that as AI agents develop further, cybercrime could reach "industrial scales."

This is an important formulation because the concern is not isolated experiments but the risk of a mass assembly line. The cheaper a single contact with a potential victim becomes, the more profitable it is for criminals to scale volume. Even if the success rate of attacks doesn't change radically, profit grows through speed, personalization, and near-continuous operation.

How Schemes Are Changing

AI doesn't necessarily invent new types of fraud—it amplifies already-known methods and makes them scalable. In practice, this means familiar threats become more convincing, faster, and cheaper to launch. Scenarios are particularly vulnerable when mass data processing is required, communication narratives must change quickly, or many parallel conversations must be managed.

• Personalized phishing emails and messages without obvious language errors
• Rapid text adaptation for a target's position, country, or specific situation
• Automated correspondence management and handling of standard objections
• Mass launch of new campaigns with low testing costs
• Reuse of successful scenarios with minimal manual refinement

This increases the burden on both regular users and organizations. Employees find it harder to distinguish fake messages from normal business communication, while security teams must respond not just to individual attacks but to the rapid pace of their updates. As attackers gain access to cheaper automation tools, the defensive side must spend more resources on training, filtering, and checking suspicious activity.

What This Means

The story of 4.5x profit growth illustrates a simple point: AI has become not just a productivity tool for business but an amplifier of criminal schemes. The faster more autonomous agents reach the market, the more critical it becomes to protect processes, maintain digital hygiene, and scrutinize non-standard requests—otherwise, fraudsters will achieve genuine economy of scale.