Artificial Intelligence (AI) is everywhere. And whether you’re on board with it or not, it’s being used to enhance, evolve, or completely transform the way we do things. Payments is no exception with a rapid adoption of AI that is reshaping payments systems from cross-border transactions to payment regulation management, and banking and payment processes.
Earlier this year, Deloitte said that 83% of financial institutions are considering AI for enhancing fraud-fighting capabilities. In 2024, the Bank of England reported that 75% of firms were already using AI, with a further 10% planning to adopt it over the next three years.
AI is still emerging technology
Despite its proliferation, AI is still widely considered to be an emerging technology. It changes daily and while many sectors are relying heavily on it, we all know that AI can (and does) sometimes get things wrong.
In payments, AI is proving to be a solution, and a problem.
- AI is being used as a solution to the growing online payments fraud issue, which cost online merchants over $43 billion in 2024. It assesses online purchases to identify the likelihood of fraud and blocks transactions that exceed specified risk thresholds.
- But, anti-fraud systems incorrectly flag and block legitimate transactions more often than not, leading to an even more significant issue known as false positives. False positives cost online merchants $443 billion in 2024. And this is just in dollars not transacted, not to mention the significant cost of reputational damage.
In 2023 J.P. Morgan reported on a case where introducing an automated anti-fraud system led to an increase in false positives because the system’s thresholds were too sensitive, wrongly rejecting legitimate card‑not‑present transactions. Despite the high risk parameters, chargebacks increased to 3% of transactions and the merchant was placed on the Visa Chargeback Monitoring program, resulting in penalty fees and increased manual overheads.
Even IBM notes that AI fraud detection can sometimes create false positives. Some may consider this an acceptable consequence of fraud prevention, but those consequences shouldn’t be underestimated – especially when they are ten times more costly than the very problem they are meant to be solving.
As a workaround, many merchants introduce more than one AI tool to detect fraud – in some cases this can be up to five. You can only imagine the cost to the merchant, let alone the volume of customer data needed to run effectively, which then puts the cardholder at greater risk if a data breach were to happen. It’s all necessary for the AI tool to run efficiently, but is this really the right approach?
The future of fraud-free online payments starts with looking back
Sometimes to move forwards, the best approach is actually to look back. The issues that plague online merchants are virtually non-existent in physical stores. The reason – chip and PIN.
Chip and PIN has proven over several decades to be the safest, most trustworthy way to accept payments. The presence of the chip and entering of the PIN provide categoric cardholder verification – known as a card-present payment. As we know, there is no chip and no PIN with payments in digital channels, hence why these are called card-not-present (CNP) transactions. And CNP are the root cause of all the issues outlined above.
CPoI® + AI: The anti-fraud power couple
Merchants want reliability, fewer disputes, and higher approval rates, all of which CPoI® delivers on its own without friction. Stolen cards and synthetic identities can’t pass a PIN check, which boosts trust and enables consistent conversion.
But, if you use CPoI® with AI, you get a powerful anti-fraud combination that results in better risk performance, zero false positives, and lower fraud losses for everyone in the payments ecosystem. The genius is in its simplicity: deploy CPoI® to verify when the AI is uncertain if a transaction is fraudulent:
- If the PIN is correct
- And the account has adequate funds
- The transaction proceeds
By requiring the customer to tap their card against their own mobile device and enter their PIN, this provides categoric verification of the cardholder’s identity. It’s a proven safeguard against fraud and places transaction liability with the card issuer (instead of the merchant) providing greater commercial sustainability. Furthermore, the data can be fed back to sharpen the AI model.
Old and new: the best of both worlds
In an industry racing toward complexity, CPoI® offers a refreshingly simple, proven solution to a very modern problem. By combining the trusted security of chip and PIN with the analytical power of AI, merchants don’t have to choose between fraud protection and customer experience. They can have both.
It’s not about replacing AI, but complementing it with a tool that’s stood the test of time. Together, CPoI® and AI don’t just reduce fraud, they restore trust in digital payments.
