AI Governance Failures Threaten Bank Returns as Singapore Lenders Scale Automation

Bank executives warned at the Asian Banking & Finance Summit on July 1 that poor AI governance — not technology limitations — poses the greatest risk to returns on AI investment in the financial sector. Citing an MIT study, former Prudential chief strategy officer Sherwin Siregar noted that 95 percent of generative AI spending currently yields no measurable business outcome, underscoring the gap between deployment and value creation.

Executives from ING, Maybank Singapore, and Green Link Digital Bank echoed the concern. ING's country manager for Singapore, Anand Sachdev, cautioned that applying AI to inefficient or broken processes does not fix them — it amplifies the damage. Maybank's Mohammed Meraj Khan observed that most organisations currently use generative AI only as a productivity assistant and that larger returns will only materialise when AI is embedded into core enterprise workflows. Green Link's David Song stressed that governance must be a continuous lifecycle, not a one-time checklist, because models can drift as real customer data flows in.

The panel converged on several requirements for safe AI scaling in banking: models must be ethical, explainable, traceable, and free from bias. Leadership must also become AI-fluent enough to decide which decisions to delegate to machines and which to keep under human judgment. Siregar framed it succinctly: AI gives speed, but speed needs direction — and direction comes from humans.

Why it matters for Singapore: Singapore's status as a regional banking hub means the governance decisions made by its lenders will set standards for financial AI across Southeast Asia. With the Monetary Authority of Singapore actively shaping AI guidelines for the sector, the industry's ability to move from experimentation to governed, value-generating deployment will determine whether Singapore's banks lead the region in AI-driven efficiency or become cautionary tales in scaling without guardrails.