A new white paper from Visa, Using AI to Secure the Future of Payments, highlights the changing characteristics of the fraud landscape, and what payment businesses need to do to stay ahead of the criminals.
And, as one might expect, Artificial Intelligence (AI) is set to play a central role – in both the threats to and protection of the payment ecosystem.
As the paper puts it: “Almost all the AI tools, tricks and capabilities available to us on the surface web already have their equivalents on the dark web – from AI-enabled marketplaces to identify and buy specialist skills, to AI-enabled dark web search engines, to AI chatbots configured specifically for crime, to polymorphic malware creation engines, to AI-powered biometric bypass and deepfake capabilities covering voice cloning, image cloning, and video cloning.”
To exacerbate matters, the sheer complexity of today’s multi-party ecosystem, with its tangled web of vendors, service providers, and intermediaries, means the attack surface is much larger, the lines of responsibility are more blurred, and the potential points of compromise are more numerous.
However, Visa points to the success of its investment in AI-enabled analytics capabilities to keep several steps ahead.
The organisation, it reports, has invested more than $10 billion in cybersecurity protection over the last five years, preventing an estimated $41 billion worth of fraudulent transactions in the past year, and reducing the fraud rate on at-risk accounts by more than 21%.
Against this backdrop, fraudsters are shifting their attention towards alternative payment methods, such as account-to-account payments and crypto payments, which are often subject to fewer safeguards, can be more prone to social engineering attacks, and typically have less mature consumer protection measures.
They are also probing for pockets of vulnerability within card-based payments – which means that, often, they identify and target issuers with lax controls or under-developed fraud detection, prediction and prevention capabilities.
In response, Visa is encouraging more banks to make fuller use of its suite of AI-enabled, scheme-agnostic tools which are offered as part of the Visa Protect range of enterprise risk solutions.
Many of these tools use advanced analytic techniques to enable issuers to spot fraud and safeguard transactions.
For example: Visa Advanced Authorisation (VAA) is an AI-powered detection engine that generates a risk score for each transaction by analysing the entire network for emerging fraud patterns and identifying transactions that don’t fit individual usage patterns.
Visa Risk Manager (VRM) is an online platform that issuers can use to manage rules for both Visa and non-Visa transactions in real-time to decline high-risk purchases and wallet provisioning requests and flag suspicious transactions that warrant further investigation.
Visa Analytics Platform (VAP) is a web-based application suite that delivers data-driven insights and benchmarks, helping issuers to track their fraud performance, compare it with industry norms, identify emerging fraud patterns, and adjust their authorisation rules accordingly.
The AI-enabled toolset can also be used to support an issuer’s authentication strategy, helping to remove roadblocks from the customer experience, and enabling them to stop fraud without having to stop the flow of business.
For example, the Visa Protect Authentication Intelligence (VPAI) service uses risk-based authentication (RBA) to evaluate ecommerce transactions in real-time using AI risk scoring models.
Drawing on hundreds of datapoints, like device IDs and IP addresses, this enables issuers to see when additional authentication is warranted, and when it’s not necessary to invoke strong customer authentication (SCA).
In a separate whitepaper on risk-based authentication, Visa reported that, by combining RBA with EMV’s 3DS, it’s possible to reduce challenge rates to less than 25% without increasing fraud.
Importantly, many of the tools that Visa has developed for card-based payments can also be deployed for other payment types. For example, the whitepaper reports on the results of a pilot with Pay.UK to analyse billions of historic UK retail bank transactions – which uncovered an additional 54% of fraud, representing a potential saving of GBP330 million.
By deploying Visa Protect for account-to-account payments, Visa says issuers can help banks to eliminate solos, boost visibility into real-time payment scams, and enable a true multi-rail fraud strategy.
Download a free copy of “Using AI to Secure the Future of Payments”, the new white paper from Visa.
The post Using AI-enabled analytics to thwart AI-enabled fraudsters appeared first on Payments Cards & Mobile.