Will we control AI or will AI control us?

But behind all the hype, hyperbole and mind-boggling valuations lies a more important question: what will work look like in the near future? What kind of people should we still be hiring? What type of skills do we need? Which jobs will disappear? That question may be even more important for the thousands of students starting a degree for a profession that may have changed beyond recognition by the time they graduate.

I immersed myself in those questions this summer. The range of answers is broad.

No one works here

At one end of the spectrum we can find Paul Cheek of MIT Sloan. His new book has a fantastic title: No One Works Here. Cheek describes the ultimate AI-native company: AI agents make more and more decisions and carry out more and more processes, meaning fewer and fewer people are needed to build ever-larger companies. Silicon Valley has been dreaming for years of the first one-person unicorn: one entrepreneur using AI to build a billion-dollar company. Cheek goes one step further: no one works here.

At the other end is Sarah O’Connor of the Financial Times. In We Are Not Machines, she travels through the history of work and technology. The Industrial Revolution made humans more productive, but reduced millions of people to human components of gigantic machines. Think of Charlie Chaplin in Modern Times: a person who no longer operates the machine, but whose body and rhythm are dictated by it.

O’Connor wonders whether we are about to make the same mistake with AI. Especially now that Google DeepMind’s Demis Hassabis predicts that AI could be ten times bigger than the Industrial Revolution and perhaps unfold ten times faster.

The Reverse Centaur

But the best book I read this summer about what work will look like in the future comes from blogger, journalist, and science fiction author Cory Doctorow. Doctorow is a delightfully blunt enfant terrible of the tech industry and a fierce critic of Big Tech. His latest book deserves an award for the title alone: The Reverse Centaur’s Guide to Life After AI.

We know the centaur from Greek mythology: human on top, horse below. It is a brilliant metaphor for technology. The human decides where to go; the horse provides speed, strength and endurance. That, in my view, is AI at its best. The doctor who uses AI to spot patterns faster, but makes the diagnosis and retains responsibility. The programmer who uses AI to write code faster, freeing up more time for architecture and quality control.

But Doctorow warns about the opposite: the reverse centaur, which turns that creature upside down: the machine is on top and the human is underneath. The human does not use the machine. The machine uses the human.

Think of a warehouse worker with a scanner. An algorithm determines where they need to go, which package they need to pick up, how many seconds they have to do it and where they need to go next. Their hands and legs execute what the software has decided. This is not a human enhanced by technology, but a human actuator for an algorithm.

Superpowers or slave to the algorithm?

We are going to see both phenomena at the same time. Some people will gain superpowers thanks to AI. One person will be able to do what once required ten. Their autonomy, creativity and impact will increase: centaurs. But others will end up in organizations where algorithms determine what they do, when they do it, how fast they do it, and whether it is good enough: reverse centaurs.

That is why we need to start thinking about this today. In our companies, in society, but above all: in education. Which academic programs will teach people how to direct, assess, question and correct technology? Which skills keep the human on top? Critical thinking. Creativity. Judgment. Empathy. Craftsmanship. The ability to formulate a problem rather than merely execute a task. But also: which courses are still preparing people for work that may soon be broken down into hundreds of small tasks prescribed by an algorithm?

I remain a strong believer in the centaur. I believe in technology that makes us stronger without taking the wheel. In AI that expands our human capabilities rather than optimizing away our humanity. But I fear that, at the same time, we will also create an enormous number of reverse centaurs.

Perhaps that will be the most important challenge of the AI era: making sure the human stays in the saddle.

This is an adapted version of a piece that first appeared in Dutch business daily De Tijd.