Stock Markets April 30, 2026 06:58 AM

Asian Banks Step Up AI Vetting as Advanced Models Raise Cybersecurity Alarm

Regional lenders tighten controls after the latest AI security model surfaces numerous vulnerabilities; regulators and CEOs flag faster, broader threat vectors

By Jordan Park
Asian Banks Step Up AI Vetting as Advanced Models Raise Cybersecurity Alarm

Major banks across Asia are intensifying scrutiny of artificial intelligence tools after the emergence of a highly capable cybersecurity AI that has reportedly identified thousands of critical software vulnerabilities. Lenders in Singapore and beyond are reinforcing assessment procedures and emphasising internal guardrails even as they seek to harness AI benefits for coding and operations.

Key Points

  • A restricted-preview cybersecurity AI, launched under Project Glasswing, has been made available to a select group of global banks and reportedly revealed "thousands" of major vulnerabilities across major operating systems and web browsers.
  • Major Singapore banks are tightening assessments and internal controls for AI tools, describing AI as a strategic asset but emphasising rigorous validation and guardrails before deployment.
  • Regulators have warned that some banks may be lagging behind AI developments, prompting lenders to accelerate their defensive measures while still pursuing operational gains from AI.

Banks across Asia, including major Singaporean lenders, have moved to tighten checks on artificial intelligence systems after the arrival of a cutting-edge defensive AI has heightened concerns that attackers could find and exploit software weaknesses more quickly and at greater scale.

The AI in question, released as a restricted preview under a project named Project Glasswing, is being positioned as an advanced model for cybersecurity tasks. It has been made available to a selection of global financial institutions: one large U.S. bank is publicly named as a launch partner, while several other international banks have access to, or are testing, the model.

According to disclosures from the AI developer, the preview stage of the model uncovered what were described as "thousands" of major vulnerabilities across "every major operating system and web browser." Those findings have amplified worries that newer-generation AI could make it easier for malicious actors to find weak points in widely used software, increasing the potential scale of cyberattacks.

Regulatory concern has followed. Australia’s prudential regulator warned recently that banks were not keeping pace with developments in AI, highlighting a potential gap between emerging tools and existing institutional controls.

Executives at regional banks acknowledge both the added risks and the potential defensive benefits of such AI. Tan Su Shan, chief executive of Singapore’s largest bank, spoke about the model after the lender reported first-quarter earnings that exceeded expectations. She said the technology "amplifies the risk" because it increases both the speed and the volume at which weaknesses can be discovered - creating a faster-moving "blast radius" for threats. At the same time, she noted that banks can use the same capabilities to accelerate their defensive work.

Tan characterised AI overall as "a net positive" for the bank, pointing to improvements in coding and operational processes that the technology can deliver.

Market reactions in the city-state reflected investor engagement with the sector. Shares in Standard Chartered rose, as did those of Oversea-Chinese Banking Corporation, DBS Group and United Overseas Bank, with intraday moves reported for each.

Oversea-Chinese Banking Corporation, Singapore’s second-largest lender, said it supports responsible AI adoption subject to strict validation. Praveen Raina, head of group operations and technology at OCBC, said all such solutions undergo rigorous assessment and validation before deployment.

United Overseas Bank, the country’s third-largest lender, described AI as a "strategic pillar" in its digital transformation. A UOB spokesperson said the bank governs AI use through established cybersecurity controls and clear internal guardrails, and that its approach to innovation is disciplined and responsible.

Standard Chartered’s chief executive, Bill Winters, called the advanced AI model a "sensational representation" of a wider trend toward heightened cyber risks, while saying the bank is well prepared. He added that the underlying cyber threat has been present for many years, but that the level of sophistication of threats has increased.

Bank leaders and corporate technology chiefs are now walking a balance between caution and adoption: strengthening vetting and internal controls while continuing to pursue AI-driven efficiencies in coding and operations. The stated intent among regional lenders is to ensure new tools are deployed only after rigorous checks and governance are in place.

At the same time, observers and regulators warn that the speed with which frontier AI can identify weaknesses may outpace some institutions' current defences unless those defences are rapidly updated and scaled.

The episode has prompted banks to intensify internal validation of AI solutions and to reaffirm existing cybersecurity frameworks as they integrate new technologies into business and operational workflows.


Market snapshot (selected Singapore-listed banks)

  • Standard Chartered: intraday rise reported.
  • Oversea-Chinese Banking Corporation: intraday rise reported.
  • DBS Group: intraday rise reported.
  • United Overseas Bank: intraday rise reported.

Concluding note

Financial institutions in Asia are tightening their controls around AI tools in response to the emergence of powerful cybersecurity models that can both expose widespread software vulnerabilities and be repurposed by attackers. Bank executives emphasise the dual-use nature of AI - a faster route to both attack and defence - and stress the need for disciplined deployment under robust cybersecurity governance.

Risks

  • Faster vulnerability discovery - Frontier AI could enable attackers to identify software weaknesses more rapidly, increasing the speed and scale of potential cyberattacks; this primarily impacts the financial sector and technology vendors.
  • Regulatory gap - Regulators have warned that banks may not be keeping pace with AI developments, creating risk that existing cybersecurity frameworks may be insufficient unless updated; this affects banks and supervisory authorities.
  • Dual-use dilemma - The same AI tools that can strengthen defensive capabilities can also be repurposed by malicious actors, complicating risk management for banks and firms deploying AI within operations and software development.

More from Stock Markets

Brockman Reveals Near-$30 Billion OpenAI Stake and Financial Links to Altman During Musk Trial May 4, 2026 California Launches Probe into Federal Deal That Scrapped Central Coast Offshore Wind Project May 4, 2026 Pilots Union Praises Kirby’s Merger Vision, Stops Short of Endorsing Deal May 4, 2026 Embraer Sees Follow-On Middle East Defense Sales After UAE C-390 Agreement May 4, 2026 Intel hires long-serving Qualcomm executive to oversee PCs and physical AI unit May 4, 2026