Think about how many times a week you swipe your card, tap your phone on an e-reader, or click “complete purchase” on a checkout page. Whether you’re buying a latte or ordering new boots, how often does it cross your mind that hackers could break into your account?

For most consumers, the answer is probably almost never. Digital commerce has become second nature, and most people don’t think twice about their financial information being unsafe.

But criminal tactics are always evolving, so detecting and preventing fraud has become an arms race. Today, criminals are using increasingly sophisticated methods to take advantage of the digital ecosystem and digital transactions. Last year, cybercrime cost an estimated $5 trillion across the globe—close to 5% of the world’s GDP.

To help fight fraud, technology companies are embracing artificial intelligence (AI). Mastercard has been at the forefront, using AI tools for more than a decade to uncover financial crime. In 2023, Mastercard’s AI-powered insights protected more than 143 billion transactions and deflected $20 billion in fraud across its network.

Staying ahead of the bad guys will require work from the entire ecosystem, so we’re sharing what we’ve learned—and forecasting how AI can reinforce our collective defenses going forward.

FINDING THE SIGNAL IN THE NOISE

Every day, humans generate 2.5 quintillion—that is, 2,500,000,000,000,000,000—bytes of data. Camouflaged within the noise are signals that can help us detect and prevent cybercrime. But the information is useless if we can’t analyze it. That’s where AI comes in. Designed to find patterns in reams of data, AI can detect subtle anomalies and keep up with fraudsters in real time. Mastercard’s AI solutions help protect customers and fight fraudsters at each stage of the transaction, from the moment a person opens their browser or app until a purchase is complete.

ARE YOU WHO YOU SAY YOU ARE?

The first step is to make sure the user is who they claim to be. Thieves bypass anti-fraud systems by masquerading as real customers. A common strategy is to steal (or buy on the black market) a consumer’s username and password and seize control of their account. The best way to reduce this type of fraud is by better safeguarding customers’ personal information. To help banks and businesses do this, Mastercard’s cyber capabilities like CyberQuant and RiskRecon diagnose an organization’s vulnerabilities and recommend the best defenses. Mastercard also provides an extra layer of protection on top of the bank’s systems: Our Safety Net service monitors every transaction on the Mastercard network and uses AI to spot emerging fraud attacks. Even before issuers and processors become aware of a threat, Safety Net helps minimize losses by blocking transactions that match the attack’s characteristics.

Mastercard AI tools also can help banks and merchants confirm a customer’s identity before they begin transacting or at the time of account opening. Our verification solutions draw links between personal attributes—such as names, phone numbers, mailing addresses, and IP numbers—to expose patterns that indicate synthetic identities or account takeover. Other Mastercard AI tools recognize the user’s devices and unique behaviors, (e.g., one user may type with two fingers, while another always holds the phone at a 70-degree angle) to distinguish legitimate returning customers from impersonators.

FINDING FRAUD BEYOND IDENTITY

Once a customer initiates a transaction, Mastercard’s Decision Intelligence solution applies thousands of data points to help banks determine whether there’s fraud. Since 2015, Decision Intelligence has been evaluating billions of transactions a year. Now, as part of the new Decision Intelligence Pro upgrade, generative AI will scan an unprecedented one trillion data points to assesses the relationships between accounts, merchants, and devices—all in less than 50 milliseconds—so banks can protect customers while still offering them the freedom to pay instantly and seamlessly. Decision Intelligence Pro has been shown to boost fraud detection rates by an average 20% and to reduce the number of false positives by more than 85%.

Fraud can also occur in the final stage of the transaction, when the goods are delivered. As online commerce grows, “first-party” or “friendly” fraud, in which cardholders challenge genuine transactions, is on the rise. Sometimes, the customer makes an honest mistake: They forgot about the purchase, or their child used the card without permission. Sometimes, however, the customer is intentionally trying to get credit for a transaction they authorized —perhaps regretting an impulse buy or simply hoping to get something for free.

To combat this trend, Mastercard is launching the First Party Trust program, an AI-powered platform where merchants can share data about the transaction to bolster the case against dishonest chargebacks. By confirming the absence of the usual hallmarks of identity theft, the First-Party Trust program can expose friendly fraud: If a customer disputes one charge out of ten to the same merchant, but their device, account name, IP address, and card number haven’t changed, it’s strong evidence that the purchase was legitimate. Rolling out this year, the program will create greater transaction transparency, optimize approval rates, and simplify the dispute process.

A SAFER FINANCIAL SYSTEM FOR ALL

During the past decade, AI has transformed how banks and businesses protect their customers, offering unprecedented opportunities to enhance security. By harnessing AI to protect against fraud and cybercrime, we can make sure consumers and businesses continue to transact with confidence. We’re pushing to deliver more insight for our customers every day—in every interaction and transaction—as we work together to build an ecosystem that is safer for everyone.