Once, prophecies were considered enigmatic visions, interpreted through symbols and metaphors. Today, these manifest themselves in the form of advanced predictive models, powered by huge amounts of data: military movements, financial flows, social media activity. These systems do not simply predict the future; often, they actively influence it.
For example, the GDELT Project continuously monitors global news in over 100 languages, identifying geopolitical and social events in real time. This database has been used to analyze and predict conflicts and political instability in different regions of the world.
The RAND Corporation, a U.S. think tank, published in 2019 the report "Extending Russia: Competing from Advantageous Ground," outlining strategies to weaken the Russian economy through sanctions and other measures. Some of these strategies appear to have materialized in subsequent years, raising questions about the interplay between forecasting and policy implementation. RAND suggests that "targeted external pressure may cause Russia to overburden its economy in non-strategic areas."
In the field of planetary defense, NASA uses systems such as Sentry, an automated program that constantly monitors near-Earth asteroids to assess impact risk over the next 100 years. For example, asteroid 2024 YR4, discovered in December 2024, initially showed a 0.3 percent probability of impact with Earth by 2032. Later observations, supported by the James Webb Space Telescope, reduced this probability, highlighting the importance of continuous monitoring and updating predictive models.
The increasing use of artificial intelligence in geopolitics raises significant questions. A 2025 RAND Corporation article points out that although AI has the potential to transform global affairs, it is critical to understand its limitations and the ethical implications of its use in policy decisions.
There is a thin line connecting the dark past, the rational present, and a future already written but still inaccessible. It is not fantasy. It is structure. MIT has shown that predictive models can anticipate the risk of war with a margin of accuracy greater than 70 percent
We picked up the quatrains of Nostradamus, not to believe them, but to see if, underneath the symbols and metaphors, there was something that makes sense today. We entrusted them to an artificial intelligence. Not because we were looking for an answer, but to see what kind of question it would ask. The result was not a truth. It was a pattern.
Some phrases, such as "from the cosmos a fireball will rise," have begun to vibrate with NASA reports of near-Earth objects. It is not prediction. It is connection.
Others, like "the realm will be marked by cruel wars, enemies within and without will arise," sound like the synopsis of a RAND document. Not just any one, but the one for 2019: Extending Russia: Competing from Advantageous Ground.
A real report. A strategy document speculating how to "overburden the Russian economy" to trigger its internal crisis. An inverted prophecy: first you write it, then you fulfill it.
In between, invisible and fundamental, predictive AI.
Not the one that completes sentences. But the one that runs in military systems, in defense laboratories, in financial databases. The one that is trained not to imagine--but to anticipate, population movements, ethnic tensions, market crashes, reactions to an attack. It is already being used.
RAND studies it. BlackRock implements it. Intelligence considers it "a decision multiplier." Some call it a digital oracle. But the oracle does not predict: it orients.
Then something happens. The words of the past, the simulations of the present, and the raw data of the world begin to synchronize.
It is not magic. It is calculation. And perhaps illusion.
But it is real.
An algorithm cross-references a quatrain on two "rocks that will collide" with a predictive model on the risk of conflict in the Pacific. Taiwan.
Result: 67% probability of escalation by 2026.
Is this a coincidence or is it a planned prophecy?
Because at a certain point it becomes difficult to distinguish between predictors and deciders.
RAND writes. AI connects. We interpret.
And perhaps at the very moment we believe a prediction, we begin to make it true.
That is the danger.
Nostradamus spoke of "a new city with poisoned waters."
AI shows us data on experimental cities where an energy design error could cause systemic disasters. MIT has already sounded the alarm: too many smart cities are born without safety plans on digital blackouts or invisible pollution. It's not mystery. It is engineering without ethics.
Once prophecies were symbols, visions to be deciphered. Today they take the form of predictive models, fed by millions of data: migration flows, ethnic tensions, stock market trends, viral disinformation. The systems that generate them are real. They are used. And they produce not just predictions, but orientations.
Some AI models predicted the collapse of Kabul days in advance. RAND, on the other hand, simulated Taiwan-US-China escalation with increasing percentages by 2026.
In this context, the quatrains of Nostradamus are not just historical curiosities. They are open linguistic architectures in which AI, trained to recognize patterns, finds echoes.
"Two great rocks will collide."
"A king will rise from the waters."
"The garden of the world will be poisoned."
These are images, but they begin to align with RAND scenarios, geopolitical maps, simulations of environmental disasters in smart cities under construction.
At some point it becomes difficult to distinguish between prediction and planning.
If AI predicts an event and the decision maker prepares for it, does it prevent it or ... trigger it?
Perhaps it is no longer a question of who writes the future.
But of how it is written: by pattern, by pattern, by self-fulfillment.
And as we try to figure out whether the prophecies were right,
the system is already turning them into operational instructions.
The distinction between prediction and planning becomes increasingly blurred. If a predictive model indicates a significant probability of an event, policy makers may be tempted to act preemptively, turning a possibility into reality. This raises the question: are we interpreting prophecies or are we unknowingly fulfilling them?
The year 2025 could be the year when prophecies, algorithms, and geopolitics merge in a massacre game where no one-neither human nor AI-really controls the endgame.
The truth? Perhaps it's hidden in a quatrain not yet deciphered ... or in the Rand Corporation's latest report, locked away in a submerged server.
Prophecies do not come true. They are fulfilled.
- 2025, between shadows and algorithms
