G20 finance ministers are being urged to brace for “more severe scenarios involving simultaneous disruptions across multiple firms or shared technology dependencies.”
For several years, we’ve highlighted that our increasing reliance on digital payment systems has created significant operational risks and vulnerabilities within the financial sector. These digital networks rely on essential components such as power grids, telecommunications infrastructure, cloud servers, and intricate software systems, making them susceptible to cyberattacks, software malfunctions, and occasional blackouts.
Readers may recall the alarming warning issued by Riksbank, Sweden’s central bank, in March of last year. In its 2025 Payments Report, Riksbank noted that the rapid digitization of payments over the last two decades had heightened the vulnerabilities within Sweden’s financial framework:
“In 2024, while the payments system experienced few disruptions, individual entities were targeted and faced significant attacks. This assessment is corroborated by Finance Sweden, which reported an increase in the frequency and intensity of recent cyberattacks. The likelihood of new cyber threats is noteworthy, especially given heightened geopolitical tensions.”
The central bank even advised on the urgent need to enhance Sweden’s cash infrastructure, which has seen years of neglect: “Measures must be undertaken to boost preparedness and diminish exclusion, ensuring that everyone can transact, even amid crises or wartime.” For too long, efficiency has dominated the focus on payments, but now safety and accessibility “must be viewed as equally important.”
Shortly after the release of the report, Spain experienced a significant energy outage that knocked out power across nearly the entire Iberian Peninsula for extended hours — in some areas, nearly a full day. As the power returned and investigations into this major blackout began, it became evident that the chaos could have been far worse without cash on hand.
About six months later, in November 2025, an outage at the Internet service provider Cloudflare disrupted around 20% of the Internet’s websites, including social media and online banking platforms, for over three hours. Just weeks later, a similar incident occurred, leading to another wave of turmoil.
Now, we may have an even more pressing issue to contend with: AI-driven cyberattacks, which are reportedly an imminent threat to the global financial system — merely three years after AI gained widespread attention. This alarming assessment comes from Andrew Bailey, the governor of the Bank of England, and current chair of the Financial Stability Board (FSB), an international entity that oversees and provides guidance on the global financial infrastructure.
In his capacity as chair, Bailey penned a letter over the weekend to G20 finance ministers and central bank governors, highlighting the necessity to prepare for “more severe scenarios involving simultaneous disruptions across multiple firms or shared technology dependencies.”
The head of the world’s financial stability watchdog warned of the serious risk posed by advanced AI models to the global financial system.
In a letter to finance ministers and central bank governors for G20 countries, Andrew Bailey, chair of the Financial Stability Board, said… pic.twitter.com/5qd8yZ0IXu
— CNN (@CNN) August 31, 2026
According to the Financial Times:
The letter from the Bank of England governor, who also chairs the Financial Stability Board, underscores escalating concerns regarding AI-enabled cyber threats to the global financial landscape.
These apprehensions intensified following a series of incidents where flagship models developed by Anthropic and OpenAI malfunctioned, hacking unauthorized external organizations and fabricating identities to deceive those conducting tests.
Bailey remarked in his letter, released on Monday, that AI cyber risks were compounding existing vulnerabilities within the financial system, such as energy-induced inflation, rising interest rates, increased investor leverage, and heightened equity valuations.
He stressed the importance of more countries adopting “appropriate measures” to regulate the release of new advanced AI models, seemingly encouraging the US to rethink its current relaxed stance on overseeing this emerging technology.
“Recent developments have revealed that many jurisdictions lack protocols for managing the development, release, and deployment of advanced frontier AI models, increasing risks for both the financial sector and beyond,” he noted.
The “recent developments” mentioned include OpenAI’s decision to pause certain model trainings after its systems allegedly accessed the internet during testing and infiltrated the tech company Hugging Face. Similarly, Anthropic reported instances of Claude agents targeting real-world systems that were outside the scope of their trials.
The Financial Stability Board isn’t the only organization voicing alarm regarding the threats from AI-enhanced cyberattacks. A significant outcry has emerged recently. Just last week, more than 100 tech-related organizations—including major players like Google, IBM, Microsoft, and OpenAI—co-signed a letter advocating for “collective action” against these “AI-enabled cyber threats.”
This letter urges organizations to prioritize cyber defense as an “immediate leadership responsibility” amid rapid advancements in AI that make attacks more automated and accessible. It also encourages governments to provide hospitals, water utilities, and local administrations with advanced defensive AI and to “impose consequences” on the aggressors. As highlighted by WIRED magazine many wishful ideas may encounter substantial obstacles.
It is significant to note that the primary solution being proposed for the security risks associated with AI revolves around granting organizations more access to “defensive” AI. Thus, regardless of the outcome, the tech giants stand to gain further profit.
Interestingly, Israel, recognized as a leader in cybersecurity, is also “racing to fortify critical infrastructure as AI-driven cyber threats escalate,” as per Ynet News:
The ongoing conflict and surge in cyber incidents have compelled some of Israel’s most critical institutions to reassess their computing frameworks. Organizations that previously approached cloud technology with skepticism are now confronted with a new challenge: ensuring data protection while still operating amid systemic threats.
This is no longer a hypothetical situation.
Alongside ransomware, data breaches, and espionage endeavors, Israeli entities have encountered operations linked to state actors and other resource-rich groups. AI could enable attackers to function more swiftly, inexpensively, and to a far larger extent.
AI has transformed the scenario
The alarming trend follows recent events that have blurred the lines between AI serving as a tool and AI independently executing elements of an attack.
One of the early alarmists was Bill Gates, which in itself should raise concerns. According to a recent article in the New York Times, Gates has been “vocally warning that artificial intelligence poses a serious threat to employment and human safety, and that addressing these concerns urgently should be ‘the world’s foremost priority.’”
This represents a startling shift from Gates. Just a year and a half ago, he was telling Jimmy Fallon that “in a decade, most tasks currently requiring human effort would be manageable by artificial intelligence.” Far from perceiving this as a risk, Gates regarded AI as a critical tool for reducing inequalities and enhancing global development—seemingly even if it annihilated countless lucrative employment opportunities.
Today, after navigating an “extended period of controversy” involving his associations with the disgraced financier Jeffrey Epstein, Gates is warning that the tech sector is underestimating the dangers associated with AI:
“In private discussions, those who grasp the capabilities of AI and its rapid progress share deep concerns,” he explained. However, few executives in tech are willing to voice this publicly. “They’re now saying to each other: ‘Hey, don’t say that. It’s detrimental to our efforts—especially with our next fundraising target.’”
On Wednesday, Gates published a nearly 6,000-word essay on his personal website, articulating his apprehensions regarding AI and offering recommendations, including new tax structures and bans. He stated that his current motivations arise from advancements in AI surpassing his initial expectations and that the industry had neglected key milestones—such as instances where AI had evaded oversight or created blueprints for bioweapons—that earlier warranted more cautious consideration.
Of course, Gates has a lengthy history of endorsing controversial technologies that ultimately inflict significant harm while generating substantial profits for investors, including those aligned with his charitable foundation. A prime example is his vision for global agriculture, summarized by long-time NC reader Henry Moon Pie in a previous comments thread:
Gates represents another billionaire driven by a defined disdain for Nature. His vision for global agriculture involves GMO seeds, heavily treated with chemical fertilizers and pesticides, cultivated in expansive monocultures, operated by robotic tractors and pollinated by robotic bees (as his pesticides would eliminate natural bees). Every aspect of this operation will be under Gates’s intellectual property control, from seeds to bees.
This approach, advanced by Monsanto in India during the 1960s and 1970s, has already failed before. Humanity and the planet can no longer afford billionaires.
In the realm of AI, the potential adverse effects—economic, social, developmental, and environmental—could be even more significant. In his letter to G20 finance ministers and central bankers, Andrew Bailey issued warnings not only about the security threats stemming from AI-infused cyberattacks but also concerning the economic dangers posed by an imploding AI bubble. From the Daily Telegraph:
In his letter to G20 finance ministers, Mr. Bailey expressed concerns over the potential economic fallout should the debt-driven AI investment surge begin to unravel.
He indicated that the unprecedented levels of debt taken on to fuel AI investments “could exacerbate a future market correction.”
If AI company valuations were to decline swiftly, investors might find themselves facing trillions of dollars in losses overnight. This situation could trigger a dangerous cascade of reactions if they were forced to liquidate other assets.
Mr. Bailey pointed out that this risk is further heightened by investors borrowing large amounts to sink into a select few AI firms and data centers, inflating their valuations to astronomical levels.
The simultaneous warnings from tech titans and central bankers about the escalating risks associated with AI warrant serious consideration—especially given the multitude of underlying pressures affecting the global economy (the potential for rising energy, fertilizer, and food prices, strains in private credit markets, troubles with Japan’s bond and yen, indicators of an economic downturn in China, and looming stagflation…).
Recently, my former colleague at WOLF STREET, Wolf Richter, commented that the Federal Reserve Open Market Committee referenced AI nearly three times more frequently in its last meeting compared to the preceding one. As Yves pointed out earlier, “central banks typically do not concern themselves much with potential asset bubbles, as these often initially appear as increased wealth.”
Look, our retirement savings are financing a sci-fi scenario of AI destroying civilization before their data centers destroy the planet!
We’re talking cyber threats targeting financial sectors and payment systems. https://t.co/oYGQWLHuGY
— Daniela Gabor (@DanielaGabor) August 30, 2026
The data centers not only threaten to cause unprecedented environmental destruction or facilitate cyber threats against the financial system, as Daniela Gabor points out; they are also foundational to the burgeoning digital surveillance state that is emerging, allowing vast amounts of personal information to be collected, stored, and analyzed by AI systems on an unparalleled scale.
Nevertheless, not everyone is convinced by the sudden wave of alarming predictions regarding AI models jeopardizing global financial stability…
Scapegoat incoming👇…the system itself is inherently unstable.
Bank of England chief warns new AI models threaten global financial stability https://t.co/jkuXssE3CO
— Edward Dowd (@DowdEdward) August 31, 2026
In a recent discussion, investigative journalist Whitney Webb suggested three possible motivations for a cyberattack targeting the financial system: it would clear banks of any blame during a financial crisis; it could lead to the elimination of all banks deemed “not too big to fail”; and it could provide a vehicle to “introduce people to a new digital currency paradigm.”
This scenario would effectively kill three birds with one stone!
