Washington: The International Monetary Fund (IMF) has raised concerns over potential systemic risks to the global financial system from advanced artificial intelligence models such as Claude Mythos, calling for urgent international cooperation to strengthen cybersecurity defences.

The warning comes weeks after Anthropic unveiled Claude Mythos on April 7, 2026, describing it as its most powerful AI model to date, particularly in identifying cybersecurity vulnerabilities. The company has restricted public access to the model, citing fears of misuse.

AI’s growing role in cybersecurity threats

In a recent blog post, the IMF highlighted how advanced AI systems could significantly lower the time and cost required to detect and exploit software vulnerabilities.

According to the IMF, such models can “dramatically reduce” the effort needed to identify weaknesses in widely used systems, making it easier for malicious actors to launch coordinated cyberattacks.

Anthropic has previously stated that Claude Mythos is capable of detecting thousands of vulnerabilities across software and operating systems—many of which may go unnoticed by human experts.

This capability, while beneficial for strengthening defences, also raises concerns that attackers could exploit the same technology to target multiple institutions simultaneously.

Risk of systemic financial disruption

The IMF warned that once vulnerabilities are identified in commonly used systems, attackers could trigger “correlated failures” across financial institutions. This could disrupt critical services such as payments, banking operations, and financial transactions on a global scale.

The report emphasised that sectors beyond finance—including energy, telecommunications and public services—could also be affected, given their reliance on interconnected digital infrastructure.

Such risks are particularly concerning in an era where financial systems are deeply integrated across borders, increasing the likelihood of cascading disruptions.

Attackers may gain upper hand

Another key concern raised by the IMF is the speed at which AI-powered cyberattacks can be executed. These systems can operate at machine speed, enabling attackers to exploit vulnerabilities faster than organisations can patch or mitigate them.

The IMF noted that this imbalance could give cybercriminals a significant advantage, especially as AI models become more sophisticated.

It also cautioned that existing cybersecurity frameworks may not be sufficient to counter next-generation AI threats, making breaches increasingly likely.

Calls for global cooperation

To address these challenges, the IMF has urged countries to collaborate more closely in developing robust cybersecurity measures. This includes sharing threat intelligence, adopting common standards, and potentially leveraging AI itself to counter cyber risks.

The organisation stressed that emerging and developing economies could be particularly vulnerable due to limited resources, making global coordination essential to prevent weak links in the system.

Limited access raises concerns

Currently, Claude Mythos is accessible only to a select group of around 40 companies under Anthropic’s Project Glasswing initiative. These reportedly include major technology firms such as Amazon, Google and Apple.

While this restricted rollout allows participating organisations to strengthen their defences, it has raised concerns about unequal access to advanced cybersecurity tools. Several banks and financial institutions outside the United States are believed to lack access, potentially increasing their vulnerability.

Countries including India, the European Union, and Canada are reportedly seeking access to the model to enhance their cybersecurity capabilities. However, there are indications that the United States may limit full access, offering only partial availability to foreign governments.

Conclusion

The IMF’s warning underscores the double-edged nature of advanced AI technologies like Claude Mythos. While they offer powerful tools to strengthen cybersecurity, they also introduce new risks that could destabilise global financial systems.

As AI continues to evolve, the need for coordinated international action becomes increasingly critical to ensure that technological advancements do not outpace the safeguards designed to contain them.