Although the volume of fraudulent transactions was down in 2022, it represented a total loss of 1.19 billion euros, according to the Banque de France. To improve data processing and limit the risk of bank fraud, banks are banking on artificial intelligence (AI).
Cheques: the most frequently defrauded means of payment
In 2022, payment fraud in France amounted to 1.19 billion euros, down 4% on the previous year, but still representing a loss of earnings for the banks.
The cheque is one of the most frequently defrauded means of payment, although its use is declining. Last year, the amount of fraudulent cheque transactions totalled 395 million euros, representing one third of the total for all payment methods combined. The main fraud continues to be the use of lost or stolen cheques, cashed by the fraudster or used as a means of payment. The average amount defrauded is 1,579 euros, according to the latest report from OSMP (Observatoire de la sécurité des moyens de paiement).
At the same time, other fraud techniques are developing, such as "rejeu" fraud, which involves intercepting a cheque as it is being sent to the beneficiary, or re-presenting an already-issued cheque forcollection.
Towards tighter controls with the help of AI
Under pressure from regulators, banks are tightening controls and " trying to improve their operational efficiency by adopting artificial intelligence ", explains Thomas Rocafull, partner at Sia Partners, interviewed by Les Echos newspaper.
More specifically, they use machine learning devices to more easily and quickly identify common fraud characteristics.
For example, Société Générale has developed a solution, "Mosaic", using machine learning and behavioral modeling algorithms to better detect fraud. Alongside the banks, specialized startups are emerging, including Bleckwen. Founded in 2019, the company has developed scoring solutions for credit organizations. Several banks are using them, including BNP Paribas, Carrefour Banque and RCI for their consumer credit offers.
" Artificial intelligence solutions applied to the fight against fraud clearly help to limit financial losses," explains François Saulnier, CEO at Bleckwen.
The prevention and detection of bank fraud should become even more effective with the emergence of generative artificial intelligence.
Tools like ChatGPT are needed to " analyze the mountains of data and sort out any anomalies ", reports Thomas Rocafull at Les Echos.
These tools, which are not intended to replace good judgment and human expertise, are being massively developed by banks in France, but also in other European countries.