The International Monetary Fund (IMF) warned that the rapid evolution of artificial intelligence technologies could pose growing threats to global financial stability, as increasingly sophisticated AI-powered cyberattacks raise the risk of widespread disruption across financial institutions and markets.
In a recent IMF blog post based on internal analytical research, the Fund said severe cyber incidents fueled by advanced AI systems could trigger liquidity pressures, undermine confidence in financial institutions, and potentially disrupt core market functions on a systemic scale.
The warning comes as global financial systems become more deeply interconnected through shared digital infrastructure, including cloud computing platforms, payment networks, software systems, and data services.
According to the IMF, advanced AI models are significantly reducing the time and cost required to identify and exploit software vulnerabilities, increasing the likelihood of simultaneous attacks targeting widely used systems.
“As AI capabilities advance, cyber risks are increasingly linked to cascading failures that may disrupt financial intermediation, payment systems, and broader market confidence,” the IMF noted.
The Fund highlighted the recent monitored release of Mythos, an advanced AI model developed by Anthropic, describing it as an example of how rapidly cyber risks could escalate.
According to IMF experts, the model demonstrated the ability to identify and exploit vulnerabilities across major operating systems and internet browsers, even when operated by non-specialists.
The development illustrates how AI-powered cyber tools could destabilize financial systems if regulators and institutions fail to adapt quickly, the IMF warned.
The report contrasted this with the cybersecurity-focused version of OpenAI’s GPT-5.5 model, which was designed under governance frameworks intended to strengthen defensive cybersecurity capabilities rather than facilitate offensive exploitation.
Still, the IMF stressed that the broader trend points toward attackers gaining a structural advantage, as automated AI systems can identify vulnerabilities and launch attacks faster than institutions can patch or mitigate them.
“In a financial system reliant on shared software and concentrated service providers, this could create synchronized vulnerabilities affecting numerous institutions simultaneously,” the analysis stated.
The IMF said the implications extend well beyond banks and financial firms because modern economies rely on overlapping digital infrastructure shared across sectors such as telecommunications, energy, and public utilities.
As a result, a major AI-driven cyberattack could spread across interconnected systems, magnifying economic disruption and increasing the likelihood of systemic crises.
The report identified several emerging concerns, warning that cyber risks are becoming increasingly systemic as faster vulnerability discovery and exploitation could transform isolated incidents into broader financial shocks.
It also noted that shared digital infrastructure across sectors such as finance, energy, telecommunications, and public utilities raises the risk of cross-sector contagion, allowing attacks targeting one industry to spread rapidly into others.
In addition, the IMF cautioned that growing dependence on a limited number of cloud providers, software platforms, and AI systems could intensify risk concentration, amplifying the impact of a single exploited vulnerability across multiple institutions simultaneously.
The IMF warned that these dynamics could produce liquidity stress, payment disruptions, market panic, and large-scale asset selloffs if multiple institutions are affected simultaneously.
For regulators and central banks, the key question is whether financial systems are sufficiently resilient to absorb cyber incidents without impairing essential financial functions.
Despite the growing risks, the IMF noted that several mitigating factors remain in place for now.
Highly advanced AI-enabled cyber capabilities are not yet widely accessible, while many financial systems continue to rely on proprietary software environments that are harder to target than open-source infrastructure.
However, the Fund cautioned that these barriers may erode quickly as AI models become more capable, more widely distributed, and increasingly vulnerable to leaks or misuse.
“Temporary containment is unlikely to substitute for sustainable defensive measures,” the report warned.
The IMF called on policymakers to adopt a resilience-focused framework for managing AI-driven cyber risks, emphasizing stronger oversight, coordinated supervision, and deeper public-private cooperation on threat intelligence and incident response.
The Fund added that financial authorities should prepare for the possibility that breaches will eventually occur despite preventive defenses.
“Resilience, especially the ability to contain attacks and recover rapidly, must become a central priority,” the IMF said.
The report concluded that while measures designed to stop attacks from spreading across systems can be costly and operationally complex, they remain among the most effective tools available for containing large-scale AI-enabled cyber incidents.
Follow the latest news on Google News
