[Paper Review] Artificial intelligence and cybersecurity in banking sector: opportunities and risks

Ana Kovačević, Sonja D. Radenković+1 more|arXiv (Cornell University)|Nov 28, 2024

Blockchain Technology Applications and Security4 citations

TL;DR

This paper examines the dual impact of artificial intelligence (AI) and machine learning (ML) in the banking sector, highlighting transformative opportunities in fraud detection and automation alongside critical cybersecurity risks such as adversarial attacks and data poisoning. It advocates for designing ML models with inherent security, trust, resilience, and robustness to ensure safe deployment in high-stakes financial environments.

ABSTRACT

The rapid advancements in artificial intelligence (AI) have presented new opportunities for enhancing efficiency and economic competitiveness across various industries, espcially in banking. Machine learning (ML), as a subset of artificial intelligence, enables systems to adapt and learn from vast datasets, revolutionizing decision-making processes, fraud detection, and customer service automation. However, these innovations also introduce new challenges, particularly in the realm of cybersecurity. Adversarial attacks, such as data poisoning and evasion attacks, represent critical threats to machine learning models, exploiting vulnerabilities to manipulate outcomes or compromise sensitive information. Furthermore, this study highlights the dual-use nature of AI tools, which can be used by malicious users. To address these challenges, the paper emphasizes the importance of developing machine learning models with key characteristics such as security, trust, resilience and robustness. These features are essential to mitigating risks and ensuring the secure deployment of AI technologies in banking sectors, where the protection of financial data is paramount. The findings underscore the urgent need for enhanced cybersecurity frameworks and continuous improvements in defensive mechanisms. By exploring both opportunities and risks, this paper aims to guide the responsible integration of AI in the banking sector, paving the way for innovation while safeguarding against emerging threats.

Motivation & Objective

• To analyze the transformative potential of AI and ML in enhancing operational efficiency and decision-making in the banking sector.
• To identify emerging cybersecurity threats linked to AI adoption, particularly adversarial attacks like data poisoning and evasion.
• To examine the dual-use nature of AI tools, where capabilities intended for security can also be exploited by malicious actors.
• To emphasize the need for AI models in banking to be inherently secure, trustworthy, resilient, and robust against manipulation.
• To guide the responsible integration of AI by proposing a framework for secure AI deployment in financial institutions.

Proposed method

• Systematic analysis of AI and ML applications in banking, focusing on fraud detection, customer service automation, and decision-making systems.
• Identification and classification of key cybersecurity threats targeting ML models, including data poisoning and evasion attacks.
• Evaluation of the dual-use potential of AI tools, where the same technology can be leveraged for both defense and cyberattacks.
• Proposed design principles for ML models emphasizing security, trust, resilience, and robustness as core system attributes.
• Integration of these principles into a risk mitigation framework for AI deployment in regulated financial environments.
• Use of case-based reasoning and threat modeling to assess real-world implications of AI vulnerabilities in banking systems.

Experimental results

Research questions

• RQ1 How do AI and machine learning enhance efficiency and competitiveness in the banking sector?
• RQ2 What are the primary cybersecurity risks associated with deploying ML models in financial institutions?
• RQ3 In what ways can AI tools be misused by adversaries to compromise banking systems?
• RQ4 What system-level characteristics (e.g., robustness, trust, resilience) are essential for securing ML models in banking?
• RQ5 How can financial institutions implement AI responsibly while mitigating emerging cyber threats?

Key findings

• AI and ML significantly improve fraud detection and customer service automation in banking by enabling real-time analysis of large-scale transaction data.
• Adversarial attacks such as data poisoning and evasion pose serious threats to ML model integrity, potentially leading to incorrect decisions or data breaches.
• The dual-use nature of AI tools means that techniques developed for security can also be weaponized by threat actors to exploit system vulnerabilities.
• ML models in banking must be designed with inherent security, trust, resilience, and robustness to withstand manipulation and maintain reliability.
• The study underscores the urgent need for enhanced cybersecurity frameworks and continuous improvement of defensive mechanisms in AI-driven financial systems.
• Without proactive design and governance, the adoption of AI in banking risks undermining data confidentiality and system integrity.