Financial firms expand AI use but limit code generation amid rising cyber risks

Financial institutions are increasing AI adoption, but they draw a line at code generation. Kaspersky’s Global AI Adoption Survey shows 59% of banking, financial services and insurance (BFSI) organisations use AI for data analysis, compared with 54% market-wide, and 54% apply it to project management versus 49% elsewhere link.

85% of BFSI respondents acknowledge AI-related risks, yet 9% believe the risks cannot be eliminated.

Shadow AI—employees using unauthorised tools such as ChatGPT—is flagged by 37% of BFSI firms, above the 30% average. This exposure bypasses traditional security controls, prompting calls for clear policies, access limits and human oversight.

Conversely, only 23% of BFSI firms employ AI for programming, an 11-point gap from the 34% market average, reflecting concerns that code errors could propagate across critical systems.

Vladislav Tushkanov of Kaspersky’s AI Technology Research Center stresses that approved LLM use must be governed, with isolation and monitoring to prevent uncontrolled shadow AI.

Overall, the sector is expanding AI for analysis and planning while reinforcing governance and traditional cyber defences to mitigate both external threats and internal misuse.