Under your leadership, national security has rightly become the defining priority of this administration. We are writing to highlight a specific dimension of that threat that has not yet received attention in the Philippines: artificial intelligence that can break into banks.
On May 13, 2026, Anthropic — a leading American AI company — gave a closed-door demonstration to the United States House Homeland Security Committee. The company showed members that its new AI model — called Mythos — could, for instance, wipe out private bank accounts. Committee Chairman Andrew Garbarino called the capability "potentially really dangerous."
On June 11, General Joshua Rudd — Director of the National Security Agency and Commander of U.S. Cyber Command — told the Senate Intelligence Committee that Mythos "broke into almost all of our classified systems, not in weeks, but in hours." The NSA's own infrastructure, among the most heavily defended in the world, could not withstand it.
The demonstration used a controlled, American-made AI model with safety guardrails. The far greater danger is what is already freely available.
China's open-source AI models — DeepSeek, Qwen, and others — are now nearly at the frontier of capability. By March 2026, Chinese open-source models surpassed all Western counterparts in global downloads. Eight of the top ten Chinese AI models are open-weight: anyone in the world can download them, run them locally, and remove their safety guardrails entirely.
Security researchers have already documented threat actors using these models to develop malware, and publicly known jailbreaking methods work against them. The U.S. Congress has called DeepSeek "a profound threat to national security." Multiple countries — Australia, Taiwan, South Korea, India — have banned these models on government devices.
The math is simple: The capability demonstrated to the U.S. Congress on May 13 will be freely downloadable — with no guardrails — within months. It will flood every market, including the Philippines.
April 2026: The U.S. Federal Reserve held emergency meetings with the CEOs of America's largest banks.
May 2026: 32 members of Congress sent a bipartisan letter urging the White House to treat AI-capable cyberattacks as a national security priority.
May 2026: New York's financial regulator issued a direct warning to all bank CISOs about "heightened cybersecurity risks posed by frontier AI models."
June 2, 2026: President Trump signed an Executive Order on AI security for critical infrastructure.
Philippine banks run on legacy technology infrastructure, most of it decades old. These systems were built before AI existed. They cannot be patched — they must be rearchitected from the ground up. This is a massive undertaking and it will take time.
The fragility is already visible. BPI, one of the country's largest banks, has suffered repeated system-wide outages affecting millions of customers — as have BDO, Metrobank, and UnionBank. If these institutions cannot maintain basic uptime on their existing systems, they are not equipped to defend against an AI that can probe their entire codebase for vulnerabilities in minutes.
The Philippines is the fourth-largest remittance corridor in the world, with $35.6 billion flowing into the country annually. Every peso that moves through the banking system moves through software that an AI — freely available, with no guardrails — can now probe for weaknesses.
The United States has begun to act. The Philippines has not.
Razon Financial Group, in partnership with leading US frontier AI companies and top Israeli and Silicon Valley cybersecurity firms, is approximately six months from deploying a purpose-built solution for the Philippines: a new bank, built from scratch, with AI security at its foundation.
This institution would not carry the vulnerabilities of legacy systems. It would use the same class of AI to defend its customers that adversaries would use to attack them. Once operational, Razon Financial Group is prepared to extend its research, threat intelligence, and security infrastructure to the broader Philippine banking sector — helping shore up the entire financial system, not just our own.
The window to act is narrow. Once these AI models reach full frontier capability in open-source form, every unprotected financial institution in the country becomes a target. The Philippines can either be ready — or it can be exposed.