Mastercard jumps into generative AI race with model it says can boost fraud detection by up to 300% – CNBC

Generate an illustration in a cartoonish style portraying the concept of an AI-powered search engine. The scene includes a large magnifying glass floating above a cityscape, with lines representing search queries radiating out from it. Around it, there should also be a few futuristic skyscrapers symbolizing different data sources with lights within them indicating search activity. Also, represent the Copilot feature as a friendly robot lending a guiding hand to the queries. On one edge of the city, depict a group of people with diverse genders and descents looking appreciatively at the magnifying glass.

Mastercard has developed its own artificial intelligence (AI) model, called Decision Intelligence Pro, to help banks detect and prevent fraudulent transactions. The AI model, built from scratch by Mastercard’s cybersecurity and anti-fraud teams, uses transformer models and data from the company’s ecosystem to assess suspicious transactions in real-time. The algorithm is trained on data from the 125 billion transactions that go through Mastercard’s network annually. Instead of relying on textual inputs, the algorithm uses a cardholder’s merchant visit history to determine the legitimacy of a transaction. The algorithm generates pathways through Mastercard’s network to find the answer in the form of a score, with higher scores indicating expected behavior and lower scores indicating potential fraud. Mastercard claims that the technology can improve fraud detection rates by 20% on average, with some cases seeing improvements of up to 300%. The company has invested over $7 billion in cybersecurity and AI technologies in the past five years. Mastercard’s algorithm has the potential to identify fraudulent patterns and predict future types of fraud within the payments ecosystem.

Full article

Leave a Reply