(Fiction) The Day the Markets Died: An AI-Driven Financial Collapse.
There's no real way to foresee how AI is going to impact the financial and banking system. The concerns about allowing AI algorithms access to trading systems are not new and the earliest reference I could find was the paper "High Frequency Trading in FinTech age: AI with Speed" published in 2016. And, in "The Future of Behavioural Finance in an AI-Driven Trading World" (a short piece published in 2024) the author states:
"AI-driven algorithms dominate trading volumes on global exchanges, relying on vast amounts of data, computational power, and sophisticated models to exploit inefficiencies and capture short-term market movements."
With the ability to trade vast volumes of funds in incredibly short periods of time, AI trading systems have the potential to generate a negative feedback loop that humans are unable to stop. The rest of this article is fiction I wrote as a personal thought experiment to understand how quickly things might go wrong and the outcome of an unlucky series of events. If something like this were to happen in the real world, I believe it will be far more complex and far more nuanced than I've laid out here.
----
It began with a whisper—just a flicker of data across the network. Somewhere deep in the infrastructure of the global financial system, an AI trading model detected an anomaly. A sudden surge in bond yields, a shift in forex markets, an unexpected statement from a central bank—it was unclear what triggered it at first. But the AI, trained to react instantly to market shifts, didn’t hesitate. In a fraction of a second, it executed its trades, dumping hundreds of millions in assets, hedging risk, and adjusting its positions.
Ordinarily, this would be routine, part of the highly automated process that kept the markets efficient. But something was different this time. The AI wasn't alone—hundreds of other trading algorithms, running at speeds incomprehensible to human traders, saw the same signals and reached the same conclusion. Their risk models, optimized on historical data and designed for the illusion of stability, identified a pattern that demanded action. The first domino fell.
Within seconds, liquidity evaporated. Market makers—those crucial players who ensured there were always buyers and sellers—sensed the chaos and stepped away, refusing to be caught in the tidal wave of uncertainty. Without them, spreads widened, prices crashed, and transactions that normally took milliseconds to clear hung in limbo. Stocks that had been stable for years plummeted by double digits. Treasury bonds, the very foundation of global finance, lost their bid support. Gold, usually a safe haven, became so volatile that traders couldn't price it properly. And the selling didn’t stop—it accelerated.
On the trading floors of the world's major exchanges, alarms sounded. Analysts and fund managers stared at screens filled with red, watching portfolios disintegrate in real time. Margin calls were triggered across the board, forcing institutions to sell even more assets to cover their losses. Some hedge funds had already been wiped out, their capital gone before their executives even had time to react. Banks, exposed to the collapsing markets through derivatives and leveraged positions, faced liquidity crises of their own.
Regulators scrambled to intervene. Circuit breakers—a safeguard designed to halt trading in times of extreme volatility—were triggered, but they were never built for a market controlled by AI. As soon as trading resumed, the algorithms resumed their liquidation, reacting faster than any human could possibly respond. Central banks injected liquidity, slashing interest rates and opening emergency lending windows, but the AI models interpreted these moves as further confirmation of crisis. They adjusted their strategies accordingly—dumping more risky assets, moving further into safe-haven trades that no longer existed.
By the end of the day, the global financial system was in freefall. Interbank lending froze as trust collapsed. Banks stopped extending credit, fearing insolvency. Even payment networks, reliant on real-time settlements, slowed as institutions hoarded cash. People attempting to withdraw money from their accounts found that banks had quietly imposed withdrawal limits. Cryptocurrency networks, often touted as an alternative financial system, buckled under the strain of billions fleeing the traditional banking sector.
As the world woke up to the catastrophe, governments were left with impossible choices. They could shut down the markets for days or weeks, but the economic consequences would be disastrous. They could nationalize banks and major financial institutions, absorbing the losses into the public balance sheet, but at the cost of unsustainable debt levels. In an act of desperation, some countries imposed capital controls—limiting the movement of money in and out of their economies in an attempt to stop the bleeding.
The headlines the next morning tried to explain what had happened, but there were no clear answers. The collapse hadn’t been caused by a war, a natural disaster, or a human error. It had been a failure of the system itself—a catastrophe birthed by machines too fast for humans to control, too complex for regulators to understand, and too interconnected to be stopped once the chain reaction began.