Artificial intelligence is forcing financial institutions to rethink the foundations of cybersecurity as increasingly sophisticated AI-generated fraud exposes the limitations of traditional security systems.
While deepfakes have dominated public attention, banking executives and cybersecurity specialists say the real challenge extends far beyond manipulated videos or cloned voices. Instead, AI is changing how financial institutions approach identity, authentication and operational resilience across their entire digital ecosystem.
The shift comes as banks globally accelerate investment in AI-powered services while simultaneously facing a new generation of cyber threats capable of producing convincing synthetic identities, forged documents, voice clones and automated fraud at unprecedented scale.
According to Deloitte, generative AI could increase banking fraud losses in the United States from US$12.3 billion in 2023 to US$40 billion by 2027, highlighting the growing financial impact of AI-enabled crime.
At the same time, identity verification research by Regula found that while 76% of organisations have deployed technology to verify whether a real person is behind an online identity check, fewer than half are confident those systems can reliably distinguish genuine users from increasingly sophisticated AI-generated impersonations.
Matthew Renirie, chief executive and co-founder of Certified AI Access, believes many organisations are still approaching AI security from the wrong perspective.
“Most organisations are asking which AI tool they should add to their existing security stack,” he says. “The better question is whether that security architecture was designed for a world where identities, documents, voices and even customer interactions can all be generated by AI.”
He says deepfake detection remains an important control but should be viewed as only one component within a broader cyber architecture.
“Financial institutions need multiple AI technologies working together across identity verification, authentication, fraud prevention and cybersecurity. No single control will solve this problem because AI is reshaping every stage of the customer journey.”
Rather than relying on one-off identity checks, banks are increasingly moving towards layered verification models that combine document authentication, biometrics, liveness detection, behavioural analytics, device intelligence, continuous risk assessment and AI-generated media detection.
The objective is not simply to detect fraud but to establish confidence that every interaction remains authentic throughout its lifecycle.
The shift is also influencing regulatory priorities.
Authorities including the UK’s Financial Conduct Authority, the Monetary Authority of Singapore and the Financial Stability Board are placing greater emphasis on operational resilience, AI governance and digital trust as artificial intelligence becomes embedded across financial services. Their focus is increasingly moving beyond responsible AI towards ensuring institutions can maintain secure and resilient operations as AI-enabled threats evolve.
South Africa is following a similar path. The Financial Sector Conduct Authority and Prudential Authority have strengthened expectations around cyber resilience and technology risk, recognising that operational resilience has become inseparable from digital resilience.
Renirie argues that AI is creating a structural change comparable to previous technology revolutions.
“For years, cybersecurity has been about protecting systems from attack. AI changes that because organisations must now prove the authenticity of the people and interactions inside those systems as well.”
He believes this represents a broader architectural shift.
“As intelligence becomes embedded across financial services, authenticity becomes another layer of cybersecurity. Institutions won’t remain resilient by continually adding point solutions,” Renirie explains. “They’ll achieve it by redesigning their cyber architecture so identity assurance, fraud prevention, AI detection and governance operate as a single integrated capability.”
That evolution, he notes, is likely to accelerate as AI becomes embedded across customer onboarding, payments, lending, compliance and customer service.
“The institutions that adapt earliest may ultimately gain more than stronger fraud protection. In an increasingly AI-mediated financial system, the ability to verify what is real could become one of banking’s most important competitive advantages,” he concludes.
