More AI, more cyber trouble for banks

India's banks and fintech firms are grappling with a new cybersecurity threat as artificial intelligence (AI) makes it easier for attackers to exploit vulnerabilities, automate attacks, and compromise credentials. As the financial sector embraces AI across lending, payments, fraud detection, and customer service, a new cybersecurity arms race emerges.

Attackers, utilizing AI models, can probe systems faster, while financial institutions employ AI to detect, predict, and respond to threats. Razorpay CEO Harshil Mathur emphasized that defenses must also become models to match attackers' speed. Cybersecurity traditionally relied on human penetration testers, but AI can now automate this process, continuously identifying weaknesses before exploitation.

BCG's latest fintech report highlights that Indian banks face 1.6 times the cyberattack intensity compared to global counterparts, with AI-related attacks considered a top concern by 76% of leaders, yet only 38% allocate more than 10% of their IT budget to cybersecurity. Cybersecurity firm CloudSEK's Rahul Sasi pointed out that attackers now target AI infrastructure, creating a new attack surface.

Meanwhile, agentic AI systems, which can access data and applications on behalf of users, pose significant risks if compromised. Identity security firm Saviynt's Nitin Varma stressed the need for visibility into AI agents' actions, applications accessed, and permissions granted, as traditional identity management methods are insufficient for managing non-human identities.