IMF boss warns global system not ready for AI cyber threats

IMF managing director Kristalina Georgieva at IMF headquarters in Washington, DC, on April 9, 2026. [AFP]

The global monetary system is not prepared to address artificial intelligence's(AI) rapidly escalating risks, the IMF Managing Director Kristalina Georgieva has warned, as Kenya's banking sector reels from billions of shillings in annual cyber losses that experts say AI will only worsen.

Georgieva's comments come a day before the International Monetary Fund and World Bank kick off their annual spring meetings in Washington, and after US regulators last week convened an emergency meeting with top bank chiefs over a powerful new AI model from Anthropic. 

"We can't — us as a world — to protect the international monetary system against massive cyber risks," she told CBS News' "Face the Nation" programme. "We are very keen to see more attention to the guardrails that are necessary to protect financial stability in a world of AI," she said, seeking global collaboration on the issue. 

Anthropic announced on April 7 that it was limiting the release of its new "Mythos" model over risks posed by its unprecedented ability to rapidly identify security vulnerabilities. The company said it was working with a consortium of major US firms to test the model, raising concerns that foreign companies may miss out on vital safety preparations. 

The warning resonates far beyond Washington. In Kenya, which is considered Africa's fintech and mobile-money powerhouse, cybercrime has already inflicted staggering damage.  

In 2024/2025, the country lost an estimated Sh29.9 billion to hackers and cyber incidents, according to the Africa Cybersecurity Report produced by the Africa Cyber Immersion Centre. 

The banking sector has been particularly hard hit. Cyber fraud cases more than doubled in 2024, with reported incidents rising from 153 to 353, while losses nearly quadrupled to Sh1.5 billion according to the Central Bank of Kenya's Financial Sector Stability Report 2025.  

The Communications Authority logged 7.9 billion cyber threats in just the first eight months of 2025 — double the total for all of 2024. 

As Kenyan financial institutions rush to adopt AI, the risks are multiplying, analysts say. 

A Central Bank of Kenya survey released in 2025 found that half of Kenyan financial institutions have already adopted AI tools, primarily for credit scoring, fraud detection and customer service. But 54 per cent of respondents identified cybersecurity vulnerabilities as a leading risk, while 52 per cent cited a shortage of skilled AI professionals. 

"You have to ask: Is your board ready? Do you have the right talent?" Georgieva said of the global challenge. "The risks have been growing exponentially." 

The CBK survey also found that 93 per cent of institutions have asked the apex bank to issue formal guidance on AI, seeking clarity on governance, risk management frameworks and incident reporting. The CBK is now finalising a Guidance Note on Artificial Intelligence. 

The Africa Cybersecurity Report noted that AI is expanding both defensive capabilities and attack sophistication. Cybercriminals are using AI-powered tools to automate intrusions, create convincing deepfakes, impersonate voices for social engineering and exploit vulnerabilities more efficiently. 

Georgieva said key financial institutions, including central banks, need to work together and be very attentive in managing the risks of cyberattacks. "Yes, this is an issue that has been addressed here in the United States, but it is an issue that easily can present itself in other parts of the world, and that is why we need people to cooperate," she said. 

Anthropic, in a blog post announcing the Mythos model, said the AI had already found thousands of high-severity vulnerabilities, including some in every major operating system and web browser. "Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely," the company warned. "The fallout, for economies, public safety, and national security could be severe." 

Georgieva struck a similarly urgent tone. "Time is not our friend on this one,” she said.