Bank of England Governor Andrew Bailey warned on August 31, 2026, that rapid advancements in artificial intelligence pose a significant risk to the stability of the global financial system, citing potential cyber disruptions that could spread across international borders. Speaking at a financial technology conference in London, Bailey emphasized that the increasing interconnectivity of financial markets, combined with the autonomous decision-making capabilities of advanced AI systems, could amplify shocks and lead to cascading failures if not properly regulated. His remarks come amid growing concern among central banks and financial regulators about the unintended consequences of deploying AI in trading, risk assessment, and fraud detection without sufficient safeguards.
The Governor highlighted that AI-driven algorithms, particularly those operating in high-frequency trading and credit scoring, could act in unpredictable ways during periods of market stress, potentially triggering flash crashes or widespread system outages. He noted that unlike human operators, AI systems may lack contextual understanding and could execute actions based on flawed data or adversarial manipulation, such as cyberattacks designed to exploit model vulnerabilities. Bailey stressed that the global nature of finance means a disruption in one jurisdiction—whether due to a compromised AI model or a coordinated cyber intrusion—could rapidly transmit to others through interconnected payment networks, clearing houses, and liquidity providers. He called for international coordination on AI safety standards, greater transparency in algorithmic systems, and stress testing that accounts for AI-specific risks.
Bailey explained that malicious actors could use AI to generate sophisticated phishing campaigns, deepfake communications, or adaptive malware capable of evading detection by traditional security tools. These AI-powered attacks could target central banks, major financial institutions, or critical infrastructure like SWIFT or real-time gross settlement systems. He warned that if such an attack succeeded in disrupting services even briefly, the loss of confidence could trigger liquidity hoarding, market panic, and a freeze in interbank lending—similar to the dynamics seen during the 2008 crisis but potentially faster and more widespread due to algorithmic speed. The Governor urged financial firms to adopt AI risk management frameworks that include continuous monitoring, human oversight, and fail-safes that can isolate malfunctioning systems.
To mitigate these risks, Bailey advocated for a precautionary approach, including mandatory impact assessments before deploying AI in systemically important functions, stricter data governance to prevent poisoning or bias, and the development of „circuit breakers” that could halt AI-driven trading during extreme volatility. He also supported the idea of an international registry for high-risk AI systems used in finance, modeled after aviation or nuclear safety reporting systems. While acknowledging AI’s potential to improve efficiency and detect fraud, Bailey insisted that innovation must not come at the cost of systemic resilience. He concluded by urging policymakers, technologists, and bankers to collaborate urgently on building guardrails that ensure AI serves as a stabilizer—not a destabilizer—of the global economy.
What specific AI applications in finance are of greatest concern to the Bank of England? The Governor expressed particular concern about AI used in autonomous trading strategies, credit risk modeling, and fraud detection systems, especially when these operate without adequate human oversight or transparency in decision-making logic.
How might a cyberattack involving AI differ from traditional financial cyber threats? AI-enabled cyberattacks could adapt in real time to defenses, generate highly convincing fake communications to trick employees, and exploit weaknesses in machine learning models themselves, making them harder to detect and more damaging than conventional attacks.
Is the Bank of England calling for a ban on AI in financial services? No, Bailey clarified that the goal is not to prohibit AI but to ensure its deployment is safe, transparent, and subject to rigorous testing—similar to how new pharmaceuticals or aviation technologies are regulated before widespread use.