The French tax authorities now use algorithms to track down tax evaders. Artificial intelligence makes it possible to study a large amount of data and thus makes the work of Bercy employees easier.

The tax authorities recovered 9 billion euros thanks to tax audits in 2019, a billion more than in 2018. This increase is notably due to tools related to artificial intelligence. Algorithms make it possible to do data mining, data exploration thanks to artificial intelligence. Algorithms capable of detecting inconsistencies in the files between income, financial transactions or lifestyle compared to household tax declarations.

>> Find all of Nathalie Levy's programs in replay and podcast here

Track fraudsters on social networks

The taxman also wants to use these tools to delve into public information posted on social networks. For example, if a fraudster declares to reside abroad but posts photos of him in France a little too frequently, the algorithm could inform Bercy agents.

In the same way, if a user sells too much business on Leboncoin or too often puts his apartment on rental on Airbnb while he declares no income, the algorithm could detect it. He then only has to warn the agents of Bercy who remain present to verify.