Artificial Intelligence and Cyber Defense System for Banking Industry: A Qualitative Study of AI Applications and Challenges

Figures & data

Figure 1. Nodes.

Table

Theme	Key codes
AI is a tool for enhancing cybersecurity	Using AI to address AI vulnerabilities AI helps solving major banking cybersecurity problems (web-based attacks, DDoS, KYC, fraud, detecting irregular behavior, phishing) AI is more efficient compared to rule-based algorithms Deep learning and reinforced learning are useful for addressing threats in real-time Cross-channel deployment
There are obstacles for using AI for improving security	Developing in-house AI tools is inefficient and creates a potential security risk Lack of compatibility between locally developed systems and third-party software Complexity Need to train employees Regulatory and compliance requirements (transparency, privacy) Socioeconomic implications
Destructive use of AI is a threat to cybersecurity of banks in Qatar	Adversarial ML Fake/biased inputs AI-powered password attacks (GAN) Using AI for breaching mobile security (voice assistant backdoors)
Employed AI algorithms have vulnerabilities	Fictitious data Chat bot privacy/leaks Accumulation of data Redundancy

The themes are analyzed in greater detail below.