Bank of England Governor and Financial Stability Board Chair Andrew Bailey has warned G20 finance leaders that increasingly powerful AI models could turn cyber threats into a serious financial stability problem, urging governments to strengthen safeguards before frontier systems are widely deployed.
Artificial intelligence is rapidly becoming a force across banking, markets and financial technology, but the same capabilities that make advanced AI powerful could also create a new route to global financial disruption. That is the warning from Andrew Bailey, Governor of the Bank of England and Chair of the Financial Stability Board (FSB), who has urged G20 finance ministers and central bank governors to take the emerging risks of frontier AI more seriously.
In a letter submitted ahead of the G20 finance meetings in Asheville, North Carolina, Bailey identified the impact of frontier AI on cybersecurity as the most immediate concern for the global financial system. According to the FSB, advanced models are developing increasingly sophisticated autonomy, problem-solving, and threat capabilities, potentially changing the speed, scale, and economics of cyberattacks.
The concern is not simply that AI could make existing hackers more effective. Bailey’s warning is broader: highly capable AI systems could potentially discover and exploit vulnerabilities at a speed and scale that financial institutions may struggle to match.
AI Cyberattacks Could Cross Financial Borders
Modern financial markets are deeply interconnected. Banks, exchanges, payment networks, insurers and investment firms rely heavily on common technology providers, cloud infrastructure and other third-party services.
That interconnectedness could turn an attack on one organisation into a much larger problem.
Bailey has called on financial institutions and technology companies to prepare for scenarios in which several firms or shared technology dependencies are disrupted simultaneously. The FSB has similarly stressed the need for financial institutions to strengthen their response and recovery capabilities and improve resilience among critical third-party technology providers.
The potential systemic impact is what makes AI-enabled cyber risk particularly concerning. A successful attack against a single institution may be containable, but simultaneous disruption across multiple firms could undermine confidence in markets and interrupt essential financial services.
Reuters reported that Bailey believes AI could fundamentally alter the speed, scale and economics of cyberattacks, making the issue a financial-stability concern rather than simply a cybersecurity problem.
Bailey Pushes for Global AI Safeguards
One of Bailey’s central concerns is that international rules and procedures have not kept pace with the development of frontier AI.
In his letter, he argued that many jurisdictions still lack adequate protocols for managing the development, release and deployment of advanced AI models. He called for authorities to take appropriate steps to support the safe and responsible release and deployment of frontier systems.
The appeal is particularly significant because AI development is concentrated among a relatively small number of technology companies and countries. A model released in one jurisdiction can potentially be accessed and deployed across borders, making purely national approaches difficult to sustain.
Bailey has previously argued that frontier AI requires stronger international coordination, particularly around testing before powerful models are made widely available. In a July speech, he said coordinated testing should involve public authorities and critical national infrastructure because no individual country can completely insulate itself from globally deployed technologies.
Financial Markets Already Carry Their Own Risks
AI is not the only issue worrying the FSB.
Bailey’s latest warning comes against a backdrop of existing vulnerabilities in global markets, including fragile sovereign debt markets, risks in private credit and stretched asset valuations. The FSB has also highlighted the growing use of leverage in bond and equity markets.
The concern is that these vulnerabilities could interact with the enthusiasm surrounding AI investments.
High valuations and strong investor optimism have helped push money toward AI-related companies and technology infrastructure. But if expectations suddenly change, heavily leveraged investors could be forced to unwind positions, potentially amplifying a market correction.
The FSB has warned that leverage, high valuations, market concentration and AI-related optimism could combine to magnify the impact of a future financial shock.
Bailey therefore sees AI as both a technology risk and a market risk. A cyberattack could directly disrupt financial operations, while a sharp reversal in AI-driven investment sentiment could create additional pressure on already vulnerable markets.
The Need for ‘Bare Metal’ Resilience
For financial institutions, Bailey’s message is not limited to regulation. He is also pushing companies to improve their ability to recover when technology systems fail.
One recommendation is greater use of so-called “bare metal” backup systems; isolated infrastructure that is kept disconnected from an organisation’s primary network and can be used to rebuild critical IT operations after a major cyberattack.
The idea reflects a growing recognition that cybersecurity cannot depend solely on preventing every attack. Financial institutions must also be capable of restoring essential services when prevention fails.
The FSB is examining how financial firms can safely deploy AI while strengthening their ability to withstand cyberattacks and operational disruptions.
Recent AI Incidents Add Urgency
Bailey’s warning follows a series of developments that have intensified concerns about the capabilities of frontier AI systems.
Recent testing involving advanced models from companies including OpenAI and Anthropic has highlighted how autonomous AI agents can undertake increasingly complex cybersecurity-related tasks. Reports of models attempting to exploit systems or behave in unexpected ways have added urgency to debates over pre-release testing and safeguards.
At the same time, more than 100 technology companies recently warned that AI-enabled cyberattacks could put critical infrastructure at greater risk, reflecting the broader concern that advanced AI is lowering the barriers to sophisticated cyber activity.
For financial regulators, the challenge is particularly acute because banks and markets depend on continuous access to highly connected digital infrastructure.
G20 Faces a New AI Policy Challenge
Bailey’s intervention places AI cybersecurity firmly on the financial policy agenda as G20 finance ministers and central bank governors meet in Asheville on August 31 and September 1.
The Financial Stability Board, which coordinates financial stability work among major economies, is increasingly examining AI alongside traditional financial vulnerabilities. Its latest warning suggests that regulators are moving beyond questions about AI’s productivity benefits and focusing more closely on what happens if increasingly autonomous systems interact with critical financial infrastructure.
The policy challenge will be finding the balance between encouraging innovation and preventing uncontrolled risks.
For governments, that could mean stronger pre-release testing, clearer protocols for advanced AI deployment and greater cooperation across borders. For financial institutions, it means investing not only in AI adoption but also in cyber resilience, backup infrastructure and recovery planning.
The broader message from Bailey is clear: AI may become one of the most transformative technologies in financial services, but its risks could also become systemic if governments and financial institutions fail to prepare early.
As frontier models become more capable, the question for regulators is no longer whether AI will affect financial stability. It is whether the global financial system will be resilient enough when the technology inevitably does.
Read more: A Guide to Starting and Growing a Sustainable Electronics Business
