SKN | BIS Warns Banks Have Minutes, Not Weeks, to Respond as AI Accelerates Cyberattacks

Key Points:

• Frontier AI can identify vulnerabilities, develop exploits and execute increasingly complex cyber operations at machine speed, compressing the window available to financial institutions.
• The BIS says banks must accelerate vulnerability management, patching, governance and incident response as attackers potentially gain an advantage from increasingly capable AI systems.
• The risk extends to crypto markets because exchanges, custodians, payment providers and banks rely on interconnected software, cloud infrastructure and third-party technology.

The Bank for International Settlements is warning that frontier artificial intelligence is fundamentally changing the economics and speed of cyberattacks, leaving banks substantially less time to identify and repair software vulnerabilities. The warning is particularly relevant to digital-asset markets, where exchanges, custodians, stablecoin issuers and payment infrastructure increasingly depend on interconnected technology and real-time systems.

AI Is Compressing the Cyberattack Windo

w

A new Financial Stability Institute paper from the BIS says advanced AI models can autonomously identify critical vulnerabilities, develop effective exploits and execute multi-step cyber operations. The result is a compressed period between vulnerability discovery and exploitation, increasing the probability that an unpatched weakness becomes an entry point for attackers.

The scale of the acceleration is becoming measurable. The Bank of England reported that frontier AI models can now complete certain cyber tasks in minutes that previously required skilled professionals hours or days. In one controlled reverse-engineering test, GPT-5.5 completed a task in 10 minutes and 22 seconds, compared with approximately 12 hours for a human expert.

That shift creates an operational problem for financial institutions. Vulnerabilities must still be verified, prioritized, assigned, repaired, tested and deployed safely. If attackers can move faster than that entire process, traditional change-management procedures can become a source of systemic exposure.

Financial Infrastructure Faces Concentration Risk

The BIS identifies third-party dependencies as another important vulnerability. Banks frequently share cloud providers, software components and technology suppliers, meaning a single weakness can potentially affect multiple institutions simultaneously. The same architecture exists across crypto markets, where exchanges, custodians, blockchain infrastructure providers and stablecoin-related platforms often depend on common technology and service providers.

This creates a distinction between institution-specific cyber risk and broader market infrastructure risk. A successful attack against an isolated company may cause localized losses, but exploitation of a widely used software component could disrupt multiple financial institutions, payment systems or digital-asset platforms at once. The IMF has similarly warned that interconnected financial infrastructure could allow AI-enabled cyber incidents to generate wider funding and solvency pressures.

Defenders Must Operate at Machine Speed

The BIS does not argue that AI only benefits attackers. The technology can also improve vulnerability discovery, threat detection and incident response. The problem is that the benefits are potentially asymmetric: attackers need to identify one exploitable weakness, while defenders must protect an entire technology environment.

Recent industry data illustrates the growing workload. The Bank of England cited a May 2026 security update in which 75 security issues across 26 publicly recorded vulnerabilities were identified, compared with a usual monthly volume of fewer than five for that provider. Mozilla also reported fixing 423 Firefox bugs in one month, approximately 20 times its normal rate.

For banks and crypto firms, the strategic response therefore extends beyond purchasing better security tools. Management teams may need to shorten approval chains, automate testing, improve continuous monitoring and establish procedures capable of authorizing emergency patches without compromising operational resilience.

Crypto Markets Face a Parallel Resilience Test

The implications for digital assets are significant because crypto markets operate around continuous settlement and globally accessible infrastructure. An outage or breach affecting an exchange, custodian, stablecoin platform or critical technology provider could interrupt trading, transfers or liquidity at a time when markets are already operating under elevated volatility.

The BIS argues that the financial sector should treat frontier-AI cyber resilience as a strategic management responsibility, with authorities reinforcing existing operational-resilience frameworks rather than necessarily creating entirely new AI-specific regimes.

Going forward, investors will increasingly need to assess not only balance sheets, liquidity and regulatory exposure, but also the resilience of the technology supporting digital-asset businesses. The critical question is becoming whether financial institutions can discover, authorize, patch and verify vulnerabilities faster than attackers can exploit them. As AI capabilities continue advancing, that response time could become an increasingly important component of financial stability.