NewsAuthoritiesFinancial crimeAI applications

North Rhine-Westphalia deploys its own AI system against financial crime

According to its own statements, the tax and finance administration in North Rhine-Westphalia is using Forensight, a proprietary AI system, to analyse digital evidence, reportedly for the first time in Germany. The source is a report by t3n.

1,000 relevant pages from 20 million

North Rhine-Westphalia deploys its own AI system against financial crime

What happened

According to t3n, the finance administration in North Rhine-Westphalia says it is using for the first time in Germany its own AI system to analyse digital evidence in investigations. The program Forensight was developed by the state office for combating financial crime together with the Fraunhofer Institute for Intelligent Analysis and Information Systems.

According to the NRW finance ministry, the system translates documents, summarises them, transcribes audio and video files, recognises persons, companies and places, and displays their connections as a network.

Background

During searches and seizures, the ministry says, millions of files regularly accumulate, and often fewer than ten percent of them are relevant to the proceedings. Manual review ties up investigators for months or years.

Figures from the pilot

In one case, investigators identified, according to the ministry, 1,000 relevant pages from a collection of 20 million within a few weeks. In further test cases, according to the ministry, roughly 60 percent more relevant documents were found than with previous methods.

Assessment

The ministry stresses that the AI does not replace investigators: the system prepares data and makes suggestions, while humans must review, assess and decide. For other authorities, the case may show how large data collections can be pre-sorted. It remains open how the system is embedded legally and under data protection rules, and how hit rates perform in further cases.

Sources

Editorially owned by Ideal Syka. Sources and method: Newsroom & method. Tips and corrections: ai@i6eal.de.

Share
← All articles

All analyses are based on i6eal's own measurements or on clearly labelled sources. Figures are snapshots and may change; corrections are disclosed transparently.