Nearly Half of FIs Tap Deep Learning Systems to Fight Fraud

Artificial intelligence (AI) has become increasingly prevalent across various industries, offering a wide range of capabilities. In the payments industry, AI has proven to be a superior tool for fraud detection, helping businesses to enhance security and safeguard their financial transactions. 

According to a recent report by PYMNTS Intelligence, AI-powered payment systems can surpass traditional tools in transaction processing speed. While conventional technology follows static rules, AI systems can select smarter transaction pathways based on speed and reliability. This results in more efficient processing, allowing businesses to maximize cash flow by handling high volumes of payments in less time.

Another significant advantage of AI in payments is its ability to adaptively detect fraud. Unlike simple pattern recognition, AI models employ deep learning (DL) techniques, enabling them to adjust and stay ahead of ever-evolving fraud tactics. Pay-by-face algorithms, for example, have an error rate of only 0.08% in recognizing legitimate users, compared to an average error rate of 4.1%.

Against this backdrop, many organizations are already leveraging AI for fraud detection, and a majority of businesses plan to implement the technology in the next two to five years.

More FIs, too, recognize machine learning (ML) and AI as the most effective technologies to combat fraud and mitigate fraud losses. Separate research from PYMNTS found that nearly 60% of FIs plan to initiate or increase their use of ML/AI models to improve existing fraud solutions in 2023, whereas only 36% had these plans in 2022.

Interest in deep learning systems is also significant, with 47% of FIs planning to initiate or increase their use of these systems to improve their anti-fraud fraud efforts in 2023, up from 26% who planned to do so last year. 

As PYMNTS recently explained, deep learning represents a subset of machine learning dedicated to instructing artificial neural networks (ANNs) with multiple layers (deep neural networks) on how to analyze patterns and make predictions from large amounts of data. These systems have been particularly effective in areas such as image identification and processing of natural language, equipping firms and FIs with the advanced capabilities and insights needed in their efforts to combat fraud.

Take cybersecurity firm Deep Instinct, for example. Earlier this year, the firm received new funding from PayPal Ventures to strengthen its threat prevention technology powered by deep learning, PYMNTS reported.

“We are excited that PayPal Ventures sees the market potential of using deep learning to safeguard against cybercriminal activity by immediately preventing even the most advanced attacks,” Deep Instinct CEO Lane Bess said at the time.