Only a weatherman knows which way the wind blows: Why science and truth matter in an AI-driven world

Last week, I watched Australian filmmaker Anthony Maras' new movie, Pressure, which dramatises President Eisenhower's January 20, 1961, response to his successor, John F Kennedy. Asked what had given the Allies a decisive advantage over the Germans on D-Day, Ike replied, 'We had better meteorologists.' The film brings to life exactly what he meant.

Well, perhaps not exactly. Eisenhower should have stated, as the movie accurately recounts, that the Allies had 'a' better meteorologist than the Nazis.

This is illustrated in the initially cool, but later increasingly fraught, exchanges between Group Captain James Stagg, a British meteorologist, and his American counterpart, Irving Krick. What appears, at first, to be little more than professional rivalry eventually reveals itself as the philosophical underpinnings of each man's convictions.


Asked to opine on whether an invasion date two days hence would prove favourable, Stagg approaches his task like a fastidious scientist, demanding the latest barometric readings from multiple weather stations and relying on his knowledge of the jet stream to prepare a comprehensive forecast.

Krick, on the other hand, an embodiment of misplaced 21st c. assurance, relies on sundry historical charts and events to extrapolate his own diametrically opposite assessment.

As things turn out, Stagg - who candidly admits to predictive uncertainty while never doubting that his is the right method--is proven correct. The rest of the film completely surrenders to the certainty of history.

Perhaps not completely, because one thing continues to hold true now as it did then. While Pressure tries to show how the greatest seaborne assault in recorded history - the Allied D-Day landings in Normandy in 1944 - could easily have failed to take place, it also reminds us not to take today's certainties for granted. History is often shaped by uncertainty, contingency, and decisions whose consequences are impossible to predict.

The business of predicting the future has always relied for its authority on the past. The practice of astrology, for instance, combined planetary and astral positions with past events to forecast the cycles of seasons, and even such black swan events as the fall of empires and the twilight of the gods.

In this, a numerical convention called mathematics developed apace, aiding and abetting the astrologer's weather or war prognosis by making it swifter, even when not always timely, more precise, even when not wholly accurate, and more acceptable, even when an inference defied common sense.

This is not to blame mathematics for errors in human judgement or understanding. As a tool, it has helped humanity make sense of the world by turning observations and experiments into measurable evidence and, eventually, universal scientific laws.

But science itself is not reductive. It is a long, often plodding, sometimes enlightened commitment, in which the scientist's success depends on patience, passion and dedication to the truth. For Stagg, meteorology was not just a means to some successful end.

He chose his pursuit, he confides to Eisenhower's secretary, because 'it is the only science where nature refuses to be conquered. You can map it, you can measure it, you can track it across oceans. But in the end, it belongs only to itself. It's pure. It doesn't care about our politics, or our wars, or our tiny human plans. It just is.'

As a first principle, then, scientific causality precedes and facilitates mathematical correlation, even when a science is not exact. Numbers, statistics, and their derivatives represent facts, not truth.

The same holds for fields such as language, human behaviour and the markets. Yet, the seminal tragedy of our times is that commercial interests have propagated the idea that knowledge, critical thinking, and creativity can be profitably surrendered to AI's proficient management of information, pattern recognition, and mathematical probability.

Ultimately, the first casualty will be the truth. But the triumph of misinformation may also result in the loss of reasoning, the scientific temper and, finally, when the cold, inductive logic of predictive nothingness prevails, the moral imperative of humanity itself.