AI can strengthen fintech cybersecurity, but risks remain — Omotola Akanni

Across the escalating cyber threats targeting financial institutions, cybersecurity and enterprise risk management professional Omotola Akanni has emphasized the growing role of artificial intelligence (AI) in enhancing cybersecurity for fintech applications. Her detailed paper, co-authored by six researchers and published in the International Journal of Computer Engineering and Technology in 2024, explores the potential of AI-driven technologies to detect, prevent, and respond to cyber threats that traditional measures struggle to counter due to their complexity.

Akanni and her co-authors note that the rapid expansion of fintech has enabled more accessible and efficient financial services, but has also broadened the attack surface for cybercriminals. They underscore AI's importance in protecting these digital platforms due to its capacity to process large volumes of data and swiftly identify suspicious activities.

Specifically, machine learning systems analyze transaction patterns and user behavior to spot activities that deviate from established norms, potentially exposing fraud, unauthorized access, or other cyberattacks.

Deep-learning models, capable of identifying intricate patterns that conventional rule-based systems might miss, further bolster fintech cybersecurity by learning from historical transaction data and distinguishing between legitimate and suspicious activities. This real-time fraud detection proves invaluable in environments where millions of transactions occur within brief periods.

Moreover, AI-powered systems can automate responses to cybersecurity incidents, such as isolating affected systems, blocking unauthorized access, and alerting security personnel, thereby reducing the time between detection and intervention.

The study also highlights behavioral biometrics—an area where AI is reshaping fintech security—as a promising development. By examining characteristics like typing patterns, mouse movements, and touchscreen interactions, AI systems create continuous user authentication profiles based on unique behavioral data, reducing reliance on traditional authentication methods like passwords.

Akanni and her co-authors acknowledge, however, that the integration of AI in cybersecurity carries significant challenges. Data privacy and ethical concerns are paramount, as AI systems need extensive datasets, necessitating strict safeguards to protect sensitive financial information and comply with privacy regulations. Furthermore, the "black box" nature of complex AI models complicates the interpretation of decision-making processes, underscoring the need for more explainable AI.

Another concern raised is the potential of AI systems to become targets of cyberattacks, such as adversarial attacks, where manipulated data deceives AI models, causing security systems to misclassify legitimate activities as threats or overlook malicious behavior. The researchers emphasize the importance of adversarial training and other defensive measures to address these vulnerabilities.

Scaling costs associated with developing and deploying advanced AI systems also pose challenges, particularly for smaller fintech companies. While cloud-based solutions offer additional computing capacity, issues around data security, privacy, and regulatory compliance arise. Akanni and her colleagues suggest that integrating AI with blockchain technology, developing explainable AI, creating personalized security systems, and fostering closer collaboration between AI systems and human cybersecurity professionals could serve as promising avenues for future development.

They advocate that AI should augment rather than replace human expertise, enabling machines to handle large-scale analysis while human professionals provide judgment and context in responding to complex threats.