Mastercard launches GPT-like AI model to help banks detect fraud
BARCELONA, SPAIN – MARCH 01: A view of the MasterCard company logo on its stand at the Mobile World Congress on March 1, 2017 in Barcelona, Spain. (Photo by Joan Cros Garcia/Corbis via Getty Images)
Joan Cross Garcia – Corbis | Corbis News | Getty Images
Payments giant Mastercard says it has built its own proprietary generative artificial intelligence model to help the thousands of banks in its network detect and resolve fraudulent transactions.
The company exclusively told CNBC that its new advanced AI module, Decision Intelligence Pro, will allow banks to better assess suspicious transactions on their network in real time and determine whether they are legitimate or not. isn’t it.
Ajay Bhalla, president of Mastercard’s cyber and information business unit, told CNBC that the new AI solution – a key component of next-generation AI – from Mastercard is a proprietary recurrent neural network that was built from the ground up by cybersecurity teams and against the company’s fraud.
“We’re using transformative models that basically help power generational AI,” Bhalla told CNBC in an exclusive interview earlier this week. “It’s all built in-house. and we have all kinds of data from the ecosystem. Because of the nature of the business we’re in, we see all transaction data that comes to us from the ecosystem.”
In some cases, Mastercard relies on open source “when necessary,” but the “vast majority” of the technology is created in-house, Bhalla said.
Mastercard’s proprietary algorithm is trained on data from the approximately 125 billion transactions that go through the company’s card network each year.
The data helps the AI to understand relationships between buyers – rather than words, as is the focus with large language models such as OpenAI’s GPT-4 and Google Gemini – and predict where fraudulent transactions occur, Mastercard said.
Heat sensor fraud patterns
Instead of text input, Mastercard’s algorithm uses a cardholder’s merchant travel history as a trigger to determine whether the business involved in the transaction is a location that the customer is likely to visit. to go
The algorithm then generates routes through the Mastercard network – a kind of heat-sensing radar – to find the answer in the form of a score.
A higher score would be one that follows the pattern of normal behavior expected of the card holder, and a lower score is out of that pattern.
This whole process takes place in just 50 milliseconds, according to Mastercard.
Bhalla said the new transaction decision technology from Mastercard will help financial institutions improve their fraud detection rates by 20%, on average. In some cases, however, the model has led to improvements in fraud detection rates as much as 300%, Bhalla said.
Mastercard says it has invested more than $7 billion in cybersecurity and AI technologies over the past five years.
That includes a number of acquisitions, including its March 2023 deal to buy Swedish cyber security firm Baffin Bay Networks.
A competitor Visa has made its own investments in AI, including a $100 million venture fund for a generational AI startup that the company founded in October 2023.
While it is still early days, Mastercard expects its algorithm to allow banks to save up to 20%, by eliminating many of the costs they would normally incur to ‘ evaluation of illegal transactions.
The real potential of Mastercard’s technology, according to Wall, is the ability to identify fraudulent patterns and trends to predict future types of fraud that are currently unknown within the payment ecosystem.
“The beauty of the Mastercard ecosystem is that we can see data from our customers around the world from these transactions,” he said. “What that does is it helps us see fraud and patterns on throughout the global ecosystem.”
Several companies in the payments and digital banking space have recently said that AI will follow
significant changes in their results. PayPal last week announced new AI-based products as well as a one-click payment feature.
Watch: Mastercard unveils Shopping Muse, an AI-powered personal sales assistant