Osegi
AI·2 min read

Forensight: NRW Deploys In-House AI to Fight Financial Crime

North Rhine-Westphalia's tax administration is using Forensight, a self-developed AI system, to sift millions of files for evidence. A pilot cut years of work to weeks.

Investigators reviewing digital evidence on screens showing a network graph of people, companies and locations.
Image: t3n Magazin
SUMODialoediviewergiftdetect

The tax administration in North Rhine-Westphalia (NRW) says it is the first in Germany to use its own AI system to analyse digital evidence in criminal investigations. The software, called Forensight, was developed by the state's Office for Combating Financial Crime together with the Fraunhofer Institute for Intelligent Analysis and Information Systems.

According to a spokesperson for the NRW Finance Ministry, the system translates documents, summarises them, transcribes audio and video files, recognises people, companies and places, and displays the links between them as a network.

Millions of files, few relevant ones

The background is a familiar problem. Searches and seizures in financial crime cases regularly yield millions of files, including documents, chats, images, audio and video recordings. The ministry says that often fewer than ten percent are relevant to the case, and reviewing them by hand can tie up investigators for months or years.

A pilot project suggests the tool can change that. In one case, investigators identified 1,000 relevant pages out of a collection of 20 million within a few weeks. Done manually, the ministry says, this would have taken years. In further test cases, investigators found around 60 percent more relevant documents than with previous methods.

Humans keep the final say

The ministry stresses that the AI does not replace investigators. Forensight prepares data and makes suggestions, while people must review, assess and decide.

Finance Minister Marcus Optendrenk (CDU) told the dpa news agency that anyone wanting to fight financial crime effectively must be as professional in analysing data as offenders are in concealing it. Forensight, he said, gives investigators a tool that filters the decisive leads from millions of records, keeping the assessment with humans while making analysis faster and more reliable.

For investigators, the practical gain is time: evidence that once took years to surface could become available in weeks, which matters in complex cases with enormous data volumes.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:forensightnrwfraunhoferfinancial crimeai investigationgermany

More from this day

See all →

More in: AI

See all →