Key Points:
- Financial firms currently adopt artificial intelligence technology at more than twice the rate of government regulators.
- Only 24 percent of financial authorities collect data on how banks use these new advanced systems.
- A massive 76 percent of the financial industry relies almost entirely on models built by OpenAI.
- Cybersecurity experts warn that new models like Anthropic’s Mythos can easily exploit weaknesses in older banking systems.
Central banks and financial watchdogs are losing the technology race. A new study reveals that government regulators simply cannot keep up with the rapid spread of artificial intelligence. Private financial institutions currently adopt advanced artificial intelligence tools at more than twice the rate of the authorities that oversee them. This massive technology gap leaves regulators blind to emerging threats in the global banking system.
The Cambridge Center for Alternative Finance published this eye-opening research on Tuesday. Researchers worked closely with the Bank for International Settlements and the International Monetary Fund to gather the information. The massive project involved surveying 350 traditional financial institutions and fintech companies. The team also interviewed over 140 artificial intelligence vendors and 130 central banks and financial authorities across 151 different countries.
The survey results show a severe lack of preparation among government watchdogs. Right now, only two out of every ten regulators claim to have advanced artificial intelligence capabilities. Furthermore, only 24 percent of the surveyed authorities actually collect data regarding how the financial industry uses these new tools. Even worse, 43 percent of the regulators admitted they have absolutely no plans to start collecting this critical data within the next two years.
Researchers warn that this massive data gap creates a dangerous blind spot. Authorities cannot successfully govern the financial sector if they navigate the risks of artificial intelligence without hard numbers. Regulators currently operate on blind optimism rather than facts. Without proper tracking, a sudden software failure or cyberattack could catch the entire global financial system completely off guard.
The release of new, highly powerful software makes this data gap even more terrifying. Earlier in April, the technology company Anthropic released an advanced model called Mythos. Cybersecurity experts quickly noted that Mythos poses severe challenges to the banking industry. Many banks still run their daily operations on legacy technology systems built decades ago. Experts worry that next-generation programs like Mythos could soon find and exploit the vulnerabilities in these older software systems on a massive scale.
Regulators across the globe have recently begun calling banks to ask how well their legacy systems handle these frontier models. Traditionally, financial watchdogs hold the banks strictly accountable for any cyberattacks or system harms. However, the report notes that enforcing this old rule becomes much harder today. Banks now buy smart, autonomous systems directly from third-party tech vendors. Because these programs make decisions without human oversight, regulators struggle to determine who actually bears responsibility when something goes wrong.
To fight this growing problem, the report offers a clear solution. Regulators must upgrade their own technology toolboxes immediately. Researchers suggest that government agencies need to adopt agentic artificial intelligence themselves. These specific programs can take actions and monitor systems automatically without a human watching a computer screen all day. Regulators need these advanced, autonomous tools just to match the speed and power of the systems they currently oversee.
Developing nations face even bigger hurdles in this technology race. Harish Natarajan, a practice manager for competitiveness and innovation at the World Bank, spoke at a recent event launching the new report. He explained that financial authorities in emerging market economies lack the basic skills and data needed to embed artificial intelligence into their daily workflows. These poorer nations risk falling completely behind as global finance becomes heavily automated.
Beyond the regulatory gap, the research team highlighted a massive concentration risk within the private sector. The global financial system relies far too much on a tiny group of powerful tech companies. The survey, which ran from October 2025 to January 2026, found that 69 percent of respondents depend heavily on OpenAI. When looking strictly at private financial firms, that reliance on OpenAI jumps to a staggering 76 percent.
Other major tech companies also hold large chunks of the financial market. At the time of the survey, just over half of the respondents used Google-built models. A little more than a third of the organizations relied on tools from Anthropic. Handing so much control to just three companies creates a highly critical third-party risk.
If OpenAI or Google experiences a major system crash, a sudden price shock, or a supply disruption, the entire global banking system could freeze. The researchers urge regulators to wake up and take these vulnerabilities seriously. Until government watchdogs gather hard data and upgrade their own software, the financial system remains highly exposed to the unpredictable risks of the artificial intelligence revolution.
