The Ministry of Finance and the Tax Agency (AEAT) have long been uploaded to the car of new technologies. In this 2021, elements of ‘big data’ have been incorporated to reinforce the surveillance of false tax residents and make macro-analysis of data that allow locating this type of tax fraud. But the Agency wants to go further and is already working on a project to tighten the fence on fraudsters with Artificial Intelligence.
Specifically, it is an initiative formulated with technological tools based on artificial intelligence to find irregular invoice issuers. In this way, the Agency wants to ‘hunt down’ taxpayers who want, fraudulently, to simulate spending to reduce the VAT quota or obtain a refund.
The tool is also used to find people who use false invoices to artificially justify subsidies. The issuance of irregular invoices also serves, according to the Agency, to organize “schemes with various companies involved that benefit from this simulated billing” or to “simulate a better income statement in order to obtain better bank financing.”
The Finance initiative generates “predictive models that facilitate the detection of issuers and recipients of irregular invoices. These models will be able to support the traditional selection system, helping to reinforce its effectiveness “.
As indicated by the Agency, “this project has worked with cases already studied by the Inspection Area in previous years. Based on all the available data, the IT tool generates a model that can be used to predict the probability of fraud in new similar cases“.
“It has been proven that companies that commit this type of fraud are grouped into a series of profiles with some common characteristics, so it is possible to generate models with a high hitting capacity“, assures the AEAT.
Of course, the body ensures that “the model developed for the detection of irregular billing does not try to replace current selection systems based on risks and on the experience of control actions, but rather complements them and facilitates their work “.
During the past year the development phase of the model of this Artificial Intelligence tool was completed, and the results of the 2018 exercise were calculated. “Its theoretical result in terms of detection capacity was high, so it was decided to move on to the exploitation phase“.
In fact, in 2021 the results have been made available to the Inspection Area for its control actions. At the same time, the model with the most available records is being retrained, and the results for the 2019 financial year are being calculated.
Throughout this year, with the results derived from the control actions, the
It will have a first assessment of the practical application of the model for its future adjustment.
Although Artificial Intelligence is a novelty in the fight against fraud, ‘big data’ is not. As Invertia already mentioned, the Tax Agency has been using tools related to this technique for a long time to locate cases of fraud.
In its strategic plan for 2021, the AEAT already indicated that this year a specific line of control has also been established based on ‘big data’ tools on taxpayers with relevant assets that ‘relocate’ their residence, pretending that they are in the foreigner when the Tax Agency understands that they actually reside in Spain. The profiles of some 126 taxpayers are being subjected to this analysis.