AI in financial crime prevention: A transformative approach

by Benjamin David, editor, TPA

Share this post

As the financial services sector confronts the escalating complexity of financial crimes, the advent of artificial intelligence (AI) is proving indispensable. Fraud, money laundering, and embezzlement have adapted to the digital age, necessitating equally sophisticated countermeasures.

In anticipation of the launch of a comprehensive AI glossary, The Payments Association’s AI and Technology sub-group is at the forefront of clarifying and standardising AI-related terminology. This article, prepared in the context of this impending launch, showcases the potential of AI-driven strategies to revolutionise financial crime prevention, illustrating the critical initiatives the subgroup aims to promote.

The growing problem of fraud

With the proliferation of digital payment methods, the last decade has seen a dramatic increase in fraud, particularly in the realm of card-not-present (CNP) transactions. These transactions, which do not require the physical presence of a card, accounted for 27% of all debit transactions in 2019 and have been increasing at a rate 10 times faster than card-present transactions.

The pressure to deliver quick and secure services to consumers who expect a seamless experience has inadvertently expanded the opportunities for fraudsters. AI plays a pivotal role in combatting this rise by analysing the vast amounts of data generated with each transaction. These data points include the identity of the payer (e.g., phone number, KYC information, card number), transaction timing, geographical details, type of expenditure, and the specific methods used for the transaction (e.g., device used, behavioural tracking). AI systems leverage this information to identify deviations from established patterns that may indicate fraudulent activity.

Dr Mark Goldspink, associate partner at The Fintech & Payments Advisory Network and member of the AI subgroup, emphasises that while current AI technologies are robust, the quality of data and a holistic approach within enterprises are critical. He states, “I believe the issues are the quality of data and a more holistic approach within the enterprise.” Goldspink adds that corporate data resilience is paramount for the future to ensure data quality, “coupled with the need to remove silos, is imperative for the future.”

Matt Clare, VP of fraud and 3D secure operations at Thredd, echoes the importance of AI, stating, “AI tools are the way forward for laser-like targeting of fraud while minimising friction to customers.” Clare highlights the necessity of utilising the significant data available to build the most effective models, including data from payment messages, card authorisations, and other events like cardholder detail changes and fraud rates on specific merchants.

Leveraging AI for detection

AI’s ability to sift through extensive transactional data and recognise anomalies is akin to finding a needle in a haystack. This capability is especially critical in environments with high volumes of transactions where manually spotting fraud would be unfeasible. AI systems are trained to detect irregularities in transaction patterns—such as unexpected changes in IP addresses or sudden alterations in IBANs associated with regular payments—that could suggest illicit activity. One of the substantial benefits of using AI in this context is its capacity to reduce false positives, which have traditionally been a significant source of financial loss for merchants, reported at approximately 10–15%. By applying additional layers of analysis to transactions flagged as suspicious, AI not only enhances the accuracy of fraud detection but also optimises the efficiency and cost-effectiveness of the operations involved.

According to Goldspink, a corporate data-resilient strategy needs to be identified early in the business development process. He suggests that synthetic data enrichment techniques could be adopted in the short term as data sets are fully established.

Clare adds, “Robust monitoring of how the data is evolving as the population grows and changes is essential, as well as tracking key metrics at as granular a level as possible, e.g., confirmed fraud vs. false positives at a product level, geographical level, transaction type, or merchant type level. This will help pinpoint where a model is becoming less effective. Refinements such as adding new data elements or events can then be assessed.”

Challenges and limitations

Despite its advantages, the deployment of AI in fraud prevention is not without challenges. The efficacy of AI systems largely depends on the availability of large, cleansed datasets. Smaller organisations, in particular, may find it difficult to access or maintain such extensive data pools, which are crucial for the effective training and operation of AI algorithms. Furthermore, major fraud detection companies often rely on aggregating anonymised data to create substantial datasets that can support deep analytical capabilities, a process that can be intricate and resource-intensive.

Additionally, as AI technology advances, so too do the techniques employed by cyber-criminals. These criminals increasingly use AI to detect vulnerabilities within systems that they can exploit. Consequently, financial institutions must continually update and refine their AI systems to stay ahead of sophisticated cyber-criminal strategies, ensuring these tools can effectively identify and react to emerging threats. For Goldspink, AI technology must be flexible enough to adopt strong continuous improvement (re-tuning) methodologies “to help fulfil the corporate data resilient strategy adopted by the organisation.”

According to Clare, “Being part of a consortium model, benefitting from the data/insights of other similar organisations, could be a solution. Care needs to be taken that the smaller organisation doesn’t experience outcomes that aren’t aligned with their customer/transaction profile if they are only a small part of a consortium model. Additionally, there needs to be a strong data strategy from day one of the business, particularly in terms of cleanliness, validation and breadth of data captured.”

Future directions

The ongoing development of AI in financial crime prevention is expected to focus on enhancing predictive capabilities and integrating more dynamic, real-time decision-making processes. As AI algorithms become more refined, they can anticipate potential fraudulent activities before they occur, thereby not just reacting to but actively preventing financial crime. Moreover, as consumer expectations continue to evolve towards faster and more user-friendly transaction processes, the role of AI will become increasingly central in balancing these demands with the need for security. Collaborations across sectors and industries, as well as partnerships between technology providers and financial institutions, will be vital in fostering an environment where AI tools can thrive and effectively counteract the sophisticated strategies employed by fraudsters.

While AI presents a promising avenue for preventing financial crime, it necessitates a balanced approach that considers both the technological advancements and the potential for misuse. With careful implementation and ongoing adaptation, AI can significantly bolster our defences against the financial crimes that have burgeoned in the digital age.

Payments Review Summer 2024
Read the entire Payments Review Summer edition here

More To Explore


Are you a member of The Payments Association?

Member benefits include free tickets, discounts to more tickets, elevated brand visibility and more. Sign in to book tickets and find out more.


Log in to access complimentary passes or discounts and access exclusive content as part of your membership. An auto-login link will be sent directly to your email.

Continue reading

AI is enhancing fraud prevention in finance by analysing transaction data to pinpoint irregularities. Subscribe to Payments Review to read the full article.

Subscribe to continue reading

Already a subscriber? Please log in to continue

Having trouble signing?

We use an auto-login link to ensure optimum security for your members hub. Simply enter your professional work e-mail address into the input area and you’ll receive a link to directly access your account.

First things first

Have you set up your Member account yet? If not, click here to do so.

Still not receiving your auto-login link?

Instead of using passwords, we e-mail you a link to log in to the site. This allows us to automatically verify you and apply member benefits based on your e-mail domain name.

Please click the button below which relates to the issue you’re having.

I didn't receive an e-mail

Tip: Check your spam

Sometimes our e-mails end up in spam. Make sure to check your spam folder for e-mails from The Payments Association

Tip: Check “other” tabs

Most modern e-mail clients now separate e-mails into different tabs. For example, Outlook has an “Other” tab, and Gmail has tabs for different types of e-mails, such as promotional.

Tip: Click the link within 60 minutes

For security reasons the link will expire after 60 minutes. Try submitting the login form again and wait a few seconds for the e-mail to arrive.

Tip: Only click once

The link will only work one time – once it’s been clicked, the link won’t log you in again. Instead, you’ll need to go back to the login screen and generate a new link.

Tip: Delete old login e-mails

Make sure you’re clicking the link on the most recent e-mail that’s been sent to you. We recommend deleting the e-mail once you’ve clicked the link.

Tip: Check your security policies

Some security systems will automatically click on links in e-mails to check for phishing, malware, viruses and other malicious threats. If these have been clicked, it won’t work when you try to click on the link.

Need to change your e-mail address?

For security reasons, e-mail address changes can only be complete by your Member Engagement Manager. Please contact the team directly for further help.

Still got a question?