BIS Warns AI in Banking Could Undermine Financial Stability and Trust

The Bank for International Settlements has issued a warning about the growing risks of AI in financial systems. Wider adoption of the technology could increase vulnerabilities in core banking operations. Consumer trust in digital security remains fragile, and failures in AI-connected systems may erode confidence further. Banks are expanding AI use across key areas such as fraud detection, underwriting, and customer support. These systems now handle increasingly autonomous decisions with minimal human input. However, as AI becomes more independent, consumers may struggle to understand the reasoning behind its choices.

Security risks extend beyond the AI models themselves. Attackers could exploit weaknesses in APIs, data sources, or interconnected services to influence outcomes. Financial institutions must secure every layer of their AI infrastructure, not just the core intelligence, to maintain trust.

The BIS highlights dependency risk as a major concern. Modern banks rely on complex, interconnected ecosystems that broaden the attack surface. Integrity risk also grows, as AI systems depend on the reliability of the data and services feeding them. To mitigate threats, banks are advised to continuously assess AI workflows, prioritise exploitability, validate runtime behaviour, and keep human oversight for critical decisions. The next phase of AI in banking will involve more autonomous systems making key decisions. Without robust safeguards, security failures could break consumer trust. Banks must address vulnerabilities at every level to prevent financial instability and protect customer confidence.