Mastercard announced CNBC the creation of your own generative AI model, Decision Intelligence Pro, this Thursday, February 1st. The objective of this system is to offer banks in its network an average improvement in the fraud detection rate of 20%, reaching up to 300% in certain cases. For partner banks, this would represent a financial gain of 20%.
Decision Intelligence Pro flags suspicious transactions to determine whether they are legitimate or not. To do this, unlike Google’s Chat GPT and Gemini, AI analyzes links from different merchants and customers to predict and detect fraud. This process is completed in just 50 milliseconds.
Instead of entering textual data, the system uses creditor history to determine whether the transaction occurs at a business where the card user is likely to shop. The AI then establishes a score. The higher the latter, the more the behavior is considered habitual. In Mastercard’s eyes, this predictive action represents the real advantage of Decision Intelligence Pro.
The AI uses an artificial neural network that was fully developed by Mastercard’s cybersecurity and anti-fraud teams. “ It is based on all types of data from internal sources. The very nature of our business means we collect a lot of transaction data “, explains Ajay Bhalla, director of cybersecurity and artificial intelligence at Mastercard.
Although the American payment systems company records around 125 billion transactions per year, AI still uses, in certain cases, open data sources.
Mastercard’s investments in cybersecurity and artificial intelligence amount to around 6.5 billion euros in the last five years. This envelope includes, in particular, company acquisitions, such as Swedish cybersecurity company Baffin Day, acquired in March 2023. Mastercard is not the only player in the payments system interest in AI. This is also the case with Visa and PayPal.