By Joe Cunningham, Head of Risk, Visa, Asia Pacific
Chatbots, virtual assistants, checkout-free shopping are all ways in which artificial intelligence (AI) has managed to make our lives easier and more efficient. Advances in computer processing and massive growth in data have given rise to a new wave of AI innovation, promising to transform industries like healthcare, transportation, retail and payments. But with that promise comes responsibility. We must ensure security frameworks are in place to better prevent and detect fraud as cyber threats evolve as quickly as the innovation does, while still delivering a seamless payment experience.
A new report by Visa, “Transforming Payment Security through Artificial Intelligence,” explains how AI is set to transform the payment security landscape. The predictive power of AI has already transformed Visa’s business, helping improve the customer experience and making us better and faster at addressing fraud and mitigating emerging security threats. In fact, Visa pioneered the use of AI in real-time fraud analytics in the early 1990s. Our most recent advancements in AI algorithms have allowed us to look at how payments can move beyond the transaction and become more automated, interactive and personalized across all channels and devices.
We use AI across more than 100 applications and capabilities at Visa, enabling us to create a more secure payments ecosystem without sacrificing the consumer experience. First, we’re using machine learning to analyze fraud migration patterns that help issuers verify card applications in real-time and at scale. Our network can flag irregularities during the application process, sending the issuer an associated risk score that they can then use to make a decision about whether to approve or decline an application.
Old-school methods of authentication no longer serve digital consumers who have come to expect a frictionless payment experience. Visa Biometrics allows consumers to authenticate themselves using their face, fingerprint or voice without going through the hassle of PINs and passwords. The solution uses machine learning to validate biometric matching and identify “spoof attacks,” adding an extra layer of security you wouldn’t find behind PINs or passwords.
False positives — when legitimate transactions are declined because of a concern about fraud – cost online merchants sales. In 2018, $278 billion in card-not-present transactions were declined globally, representing a 27% year-over-year growth. With the help of AI and sophisticated risk detection technologies, we can analyze large amounts of transaction data that flow through our network to help identify criminal activity more accurately and minimize false positives. In 2018 alone, we prevented an estimated $25 billion in fraud.
Beyond fraud detection and cyber defense, we are also using AI to power advances in consumer-facing technologies such as voice assistants, chat platforms and loyalty programs. Along with our clients, we are experimenting with AI-powered commerce experiences that anticipate and fulfill consumer needs in novel ways. This would not be possible without the confluence of trends outlined earlier, which has awarded us an exciting opportunity to change the way we pay and get paid while making payments more secure than ever. Still, the cybercrime business is estimated at $600 billion and it’s a reality we face every day. By embedding security and privacy into our products and services from the outset and using AI’s ability to decode complex patterns of data to outsmart cybercriminals, we can continue to deliver a rich, uninterrupted payment experience we all rely on and expect.
 Excludes insufficient funds and issuer/switch inoperative declines. eCommerce purchases for the year ending in CY18. YoY growth based on CY18 vs. CY17. Sales based on VisaNet authorization data. Fraud based on issuer reported TC40 (including transactions which were not processed on VisaNet.)
 PYMNTS.com, ‘Visa Advanced Authorization Blocks $25 Billion In Fraud’, June 2019
 BusinessWire, New Global Cybersecurity Report Reveals Cybercrime Takes Almost $600 Billion Toll on Global Economy, February 2018