Banking Fraud to Reach 2.2 billion Transactions by 2031

Fraudulent activities within digital banking and monetary transfers are projected to exceed 2.2 billion transactions globally by 2031, rising from 773.7 million in 2025, according to an analysis by Juniper Research. This represents a substantial growth rate of over 180%.

The study indicates that the rise of AI-enabled attacks is fundamentally altering the economics of fraudulent activity. By reducing the costs associated with generating large-scale, personalized social engineering campaigns, AI is increasing the vulnerability of banking customers. This marks a shift in focus from compromising core banking infrastructure toward targeting individual users and authenticated payment processes through social engineering.

"Alongside a decreased cost for committing fraud, AI permits bad actors to generate and refine attacks more rapidly, while instant payment systems reduce the timeframe available for banks to detect irregular behavior. Banks must, therefore, integrate behavioral, identity, and payment intelligence before authorization, rather than relying on reactive controls," noted report author Shane O’Sullivan.

The report details how generative AI facilitates more adaptive scam strategies, while emerging agentic AI allows fraudsters to coordinate complex, multi-stage attacks and adjust tactics in real-time. This increases the likelihood of fraud becoming more persistent, personalized, and challenging to detect.

Consequently, financial institutions must advance beyond simple detection of suspicious transactions toward a comprehensive understanding of evolving risks throughout the customer journey. The report suggests that banks utilize AI to establish a continuous perspective on customer risk, analyzing how behavioral and identity signals interact to permit intervention before a fraudulent transaction is authorized.