Banking’s Growing Reliance on AI
AI adoption across financial services has jumped sharply, with 86% of firms now using AI, up from 58% in 2022. In banking, common uses include decision support systems (reported by 80% of respondents), operations and IT automation (31%), and experimentation with generative AI features (56%). These trends reflect a shift from pilot projects to mainstream use in credit decisioning, transaction monitoring, customer service and infrastructure optimization.
Top Concerns: Cyber and Third-Party Vulnerabilities
Regulators and firms identify cyber threats and concentration in third-party suppliers as the dominant risks. Around 79% of respondents flagged third-party concentration as a primary concern. A separate finding indicates 80% of firms expect cyber actors to exploit AI-related vulnerabilities within one to three years. Key risk vectors include model poisoning, data exfiltration from shared platforms, supply-chain compromise of cloud and model providers, and reduced transparency from black-box models that complicate incident response.
Regulatory View and Forward Action
The Bank of England and the Financial Conduct Authority find that most firms have AI policies on paper, with 94% reporting some form of policy. However, only about half run periodic reviews or formalised testing regimes. Regulators are focused on governance gaps, third-party oversight, operational resilience and explainability.
For banks this means practical priorities: tighten third-party risk management through inventory and concentration limits, harden cyber controls around model and data pipelines, formalise testing and validation cycles, and set clear governance for generative AI use. Regular red-teaming, contractual security requirements, and incident playbooks for AI failures should be part of operational resilience plans.
As AI becomes embedded across banking, firms that treat governance and supply-chain risk as strategic issues will be better positioned to manage regulatory scrutiny and systemic threats.
