AI Might Save the World, But Only If Humans Get Out of the Way

On the evening of April 14, 1912, the RMS Titanic was sailing at full speed through the North Atlantic.

Thirty-seven seconds after the iceberg was spotted, she struck it. Two hours and 40 minutes later, she was gone. More than 1,500 people died.

No matter how many times I revisit the story, my conclusion remains the same: Captain Edward Smith bears primary responsibility for the many reasons and decisions that led to that outcome.

The Titanic question: would an AI assistant have changed the outcome?

It raises an interesting question. If, in that moment, Smith had access to an AI assistant like ChatGPT or Claude, would things have turned out differently?

It's tempting to say yes. Today's systems can synthesize vast amounts of information, identify patterns, and model outcomes in seconds.

Faced with ice warnings, speed, visibility, and trajectory, an AI system might well have recommended a simple course of action: slow down or divert.

But I suspect the outcome would have been the same.

Because Smith already had the critical inputs. Ice warnings had been received. The risks were not unknown. The failure was not informational. It was human.

We are emotional creatures who occasionally manage to be rational

Here is the uncomfortable truth: we are not rational creatures who occasionally become emotional. We are emotional creatures who occasionally manage to be rational.

That truth is not entirely a flaw. It is also what makes us human.

There is no clear logic to why we fall in love, or why certain pieces of art move us while others leave us cold. Emotion gives color, meaning, and texture to our lives.

But in moments of pressure, when reputation, experience, and instinct collide, those same emotions distort judgment. We anchor to past success. We defer to authority. We overestimate our control.

And, like Smith, we push forward when caution would serve us better.

From win-win to zero-sum: the data is not what is missing

So is it fair to draw a line from 1912 to today? I think it is.

We are now surrounded by tools that can augment decision-making at the highest levels, tools that can model complex systems, simulate scenarios, and challenge assumptions. And yet, many of the decisions shaping our world still feel reactive, adversarial, and short-term.

For the first 40 years of my life, I was brought up in a world where most leaders believed in win-win scenarios. That belief manifested in the creation of the European Union, global trade agreements, open borders, and a genuine push for companies to think as much about purpose as profit.

Those beliefs are now being replaced by the zero-sum model, where there can only be winners and losers. Ego has taken over.

Not because the data is missing. But because the humans interpreting it are.

The paradox: AI improves decisions and can manipulate the deciders

Which leaves us with a paradox.

AI has the potential to dramatically improve decision-making. Not because it is perfect, but because it is, in many contexts, less burdened by ego, fear, and bias than we are.

At the same time, it introduces a new and profound risk.

If human decisions are driven largely by emotion, what happens when we deploy systems that can understand, and increasingly influence, those emotions at scale?

We have already seen a glimpse of this through social media. Algorithms have shaped what we see, what we believe, and how we engage with the world. AI will amplify that capability significantly.

Can AI Help Us Make Better Decisions?

The question is no longer whether AI can help us make better decisions. It is whether we are willing to let it.

Because the greatest risk may not be that AI gets it wrong. It may be that we ignore it when it gets it right, or worse, that we use it to justify the decisions we wanted to make all along.

Will AI Actually Save the World?

AI might yet save the world. The capability is real, and advancing quickly.

But only if the humans holding it are willing, at least occasionally, to get out of the way.