And if central banks weren’t already temples of financial dogma, now they want to become AI labs too. Monetary orthodoxy is rubbing shoulders with algorithms, and institutions like the Federal Reserve, the European Central Bank, and even Mexico’s central bank are now giving artificial intelligence a seat at the table of economic policymaking.
The new oracle is neither gold nor the dollar — it’s machine learning.
According to a report by Reforma, more than 50 central banks are already using AI tools for tasks ranging from fraud detection and financial system monitoring to risk forecasting, text analysis, and economic modeling. The International Monetary Fund confirms it: AI is no longer an experimental luxury — it’s a tool of everyday governance.
But what kind of automated decisions are these institutions actually making? Do we really want an algorithm fine-tuning the interest rates that shape inflation, credit, and unemployment?
The image is grotesque — but real. The Bank of England’s Monetary Policy Committee already uses large language models like GPT to draft speeches, analyze meeting minutes, and synthesize economic trends. Will there come a time when the official press release isn’t signed by the governor, but by a prompt?
AI doesn’t just predict — it spies, tracks, monitors, and compares. The ECB uses it to automatically assess risk in bank portfolios. In Hong Kong, it’s used to detect unusual financial activity. In Canada, to model inflation and inflation expectations. The explosion of data and variables has made it impossible for any human team to handle everything in real time. But data isn’t wisdom.
A system trained on historical bias or shaped by political pressure can end up reinforcing decisions that disproportionately affect specific populations or sectors. What happens if the algorithm decides rate hikes don’t hurt the average citizen but do “correct” the markets? Or if it concludes that an emerging economy must endure harsher austerity to appease investors?
The question isn’t whether AI can help, but who programs it, with what data, and for whose interests.
In Mexico’s case, Banxico has begun exploring AI models to enhance its analytical and statistical work. And, as is typical, it has shown “institutional caution” — a euphemism that, in plain Spanish, means “we’re late, but don’t say it out loud.” In partnership with Columbia University, Mexico’s central bank has already tested systems for classifying economic news and analyzing sentiment in financial texts.
It may sound like good news, but it also raises concerns. Will these systems replace critical reading of the economic context? Will the algorithms be trained on café economists’ op-eds, political statements, or forecasts debunked a month later? Who watches the watcher?
A Silent Revolution — With No Regulation
More than half the world’s central banks have adopted some form of AI. But only a handful have regulatory frameworks to ensure ethical use, transparency, or oversight. In other words: the guardians of money are playing with algorithmic fire — and no extinguisher in sight.
And this isn’t hyperbole. The Bank for International Settlements (BIS) has already warned: the lack of clear rules could lead to dangerous dependencies, amplified errors, or sophisticated manipulation. Put simply, AI could become a systemic risk if its use isn’t regulated as strictly as capital flows.
Artificial intelligence is here to stay. But not all intelligence is wisdom, and not every predictive system gets it right. Central banks are diving into an ocean of data — but swimming in numbers doesn’t guarantee they’ll reach the shore of sound judgment.
“The one who controls the algorithms controls perception; and the one who controls perception manipulates the economy,” said an old anonymous banker at a Davos forum. But today, even that’s unnecessary: all it takes is for a model to say “adjust,” and the entire system responds.
Are we standing at the gates of a new financial technocracy? Or worse: of an enlightened despotism powered by artificial intelligence?
“Central banks now believe in AI. But AI doesn’t believe in central banks.”
The Era of Monetary Policy Written by Algorithms, Not Though
There was a time when central banks were seen as the last bastions of human reasoning — places of complex, nuanced decisions grounded not only in data but also in intuition, experience, and judgment. That era is over. Welcome to the century where speeches are written by robots, and financial stability is measured in “sentiment” as detected by neural networks.
Today, banks in England, Canada, Japan, and Mexico no longer read Keynes or Milton Friedman. They read prompts. And the technocratic cult has bowed before the new god: Artificial Intelligence.
According to Reforma, over 50 central banks are using AI for tasks ranging from analyzing meeting transcripts and public speeches to detecting fraud and forecasting crises. The Fed, the ECB, the Bank of Japan, and Banxico don’t just use AI — they’re institutionalizing it.
The twist? AI is no longer a “lab experiment.” It’s now another voice — silent but dominant — at the table where decisions are made about the value of your money, your credit, and your savings.
Examples are unsettling:
- The Bank of England drafts internal documents and public speeches using GPT models.
- The European Central Bank mines text to gauge market sentiment.
- Canada models future inflation using machine learning.
- Mexico, with Columbia University’s help, trains algorithms to classify economic news and assess its impact.
Yes, the same AI that recommends you cat videos now tells you whether there’s going to be a recession or an interest rate hike.
The argument is seductive: AI can process millions of data points that no human eye could ever take in. But that doesn’t make it wise. Or fair. A poorly trained, biased, or manipulated AI can recommend monetary policy as disastrous as that of an incompetent banker — just with better grammar.
The BIS has already pointed out the obvious: “indiscriminate use of AI can amplify errors, generate systemic risk, and blur lines of accountability.” Who decides whether a forecast was accurate? Who pays the price if the algorithm fails? Who audits the data bias?
Spoiler: no one. Because technological dogma has also replaced self-criticism.
In Mexico’s case, Banxico is flirting with AI, but from its usual spot: institutional laggard. It doesn’t lead — it observes. It doesn’t innovate — it imitates. And as always, it shows up late to the party… and without a gift.
What does this translate to? Models that scan press headlines, evaluate market sentiment, and predict inflation based on emotional “big data.” A modern oxymoron.
But if the algorithm says “the market is nervous,” who’s nervous? The algorithm? The bankers? Or us — the humans on the other side of the counter?
The paradox here isn’t that central banks are using AI. The paradox is that no one is regulating it. More than half of the world’s monetary authorities are already adopting these technologies, but only a few have ethical frameworks, oversight mechanisms, or error protocols in place.
In short: the architects of money are building a new system with no blueprint, no rules, and no supervisor. One machine can watch another, yes — but no machine can carry moral responsibility.
To sum it up: we’re delegating power without delegating accountability.
English 
Spanish 