Cybercriminals are increasingly using artificial intelligence (AI) tools to accelerate cyberattacks against financial institutions, with recent incidents affecting banks and financial companies across multiple countries and resulting in significant data exposure and financial losses.
South Korea has emerged as one of the latest targets of AI-assisted hacking campaigns. Attackers reportedly used an open-source AI penetration-testing tool called ARTEX, alongside AI models and coding assistants including DeepSeek and Anthropic’s Claude Code, to target financial institutions.
The attacks initially affected seven South Korean financial companies, exposing personal information belonging to approximately 68,000 people. However, the investigation expanded as authorities identified additional incidents. By October 14, South Korean financial regulators had identified security breaches affecting 16 financial companies, with approximately 100,000 reported data leaks.
According to cybersecurity researchers, attackers used AI-assisted reconnaissance to identify weaknesses in exposed systems, automate parts of the intrusion process and access sensitive customer information. The compromised systems included loan inquiry services used by financial brokers and internal mobile applications designed for bank employees.
Investigators believe the attacker may have been financially motivated, although authorities have not publicly established a definitive identity or affiliation. The incidents have also raised concerns about inadequate access controls, unpatched vulnerabilities and weak security practices across financial institutions.
A separate incident in Italy showed how AI can enable financial fraud without directly compromising a bank’s computer systems. Fraudsters reportedly used an AI-generated voice to impersonate a senior law firm partner and support a fraudulent payment request targeting Fideuram, the private banking arm of Intesa Sanpaolo.
The attackers first sent a WhatsApp message impersonating Intesa Sanpaolo CEO Carlo Messina before following up with a phone call using a cloned voice. The convincing impersonation led Fideuram’s then-chairman, Paolo Molesini, to authorize international transfers worth approximately €95 million.
Authorities and financial institutions subsequently recovered around €53 million through international cooperation involving China, Portugal and Italy. However, approximately €36 million remains missing after the funds moved through overseas accounts and were converted into cryptocurrency.
India has also reported significant financial losses linked to AI-assisted cybercrime. According to a report by The Economic Times, smaller non-banking financial companies (NBFCs) lost an estimated ₹300 crore over approximately two months, with nearly 150 complaints or police reports filed. The incidents have raised concerns about attackers exploiting weaknesses in financial technology infrastructure and payment systems.
Meanwhile, cryptocurrency exchange Bitget suffered a major security breach in September 2026. The exchange initially estimated losses at approximately $351.6 million, but later revised the figure to around $387.5 million following additional blockchain investigations. Bitget suspects the operation was consistent with techniques used by North Korean hacking groups, although attribution has not been independently established.
The growing number of incidents highlights how AI is changing the economics of cybercrime. Attackers can use AI assistants to speed up vulnerability discovery, analyze stolen information, automate reconnaissance and develop convincing impersonation campaigns. However, AI is not necessarily acting independently; human operators remain involved in directing many of these attacks.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
Financial institutions are now facing increasing pressure to strengthen their cybersecurity infrastructure, improve identity verification, enforce multi-factor authentication and regularly patch exposed systems. Security experts and regulators are also emphasizing basic security controls, particularly for third-party services and applications that connect to sensitive financial data.
