News
17/11/2025

ATTENTION!: AI uncovers patterns of illegal trading activities

The German-Singaporean research project ATTENTION! has developed an approach that makes it easier to recognise illegal trade activities such as smuggling, counterfeiting or money laundering. Explainable, graph-based AI models can identify suspicious patterns in international trade flows and provide clues to hidden networks.

Global trade flows and AI
Global trade flows - and AI that recognises suspicious patterns
© Pixabay / Pete Linforth
Global trade flows and AI

Illegal trade practices cause considerable economic damage worldwide. In Europe, smuggling and counterfeiting are estimated to destroy around 470,000 jobs every year and deprive legal companies of billions. At the same time, authorities and companies are faced with the challenge that the enormous volume of trade can hardly be managed with the available control capacities. Although individual consignments are often apprehended, the criminal structures behind them remain unrecognised.

How AI makes illegal trading patterns visible

ATTENTION! has created a data-driven understanding of how such patterns in trade can be scientifically identified. To do this, the project integrated data from numerous sources - including company data, international import/export databases, websites and customer reviews - and analysed them using robust, explainable AI models. The graph-based analysis allows conspicuous trading patterns and entire networks to be visualised so that illegal activities can be investigated holistically.

This was supplemented by a user-friendly cloud interface that allows companies to be checked. The key success of the project was a prototype from which market-ready cloud software can be developed. It is designed to automatically analyse traders, visualise trading networks and thus make the fight against illegal activities more effective - while at the same time reducing personnel costs.

From research to a market-ready solution

After the end of the project, the companies involved plan to further develop the software prototype and are looking for investors and pilot customers. Software and analysis companies have already been able to integrate findings from user tests and the handling of data into their work. The participating universities have also carried out pioneering methodological work in the AI-supported analysis of complex trading data, which they would like to continue in further research projects.

More information