This summer, I found myself battling through traffic in the sweltering streets of Marseille. At a crossing, my friend in the passenger seat told me to turn right toward a spot known for its fish soup. But the navigation app Waze instructed us to go straight. Tired, and with the Renault feeling like a sauna on wheels, I followed Waze’s advice. Moments later, we were stuck at a construction site.
A trivial moment, maybe. But one that captures perhaps the defining question of our era, in which technology touches nearly every aspect of our lives: who do we trust more – other human beings and our own instincts, or the machine?
The German philosopher Immanuel Kant famously defined the Enlightenment as “man’s emergence from his self-imposed immaturity.” Immaturity, he wrote, “is the inability to use one’s understanding without guidance from another”. For centuries, that “other” directing human thought and life was often the priest, the monarch, or the feudal lord – the ones claiming to act as God’s voice on Earth. In trying to understand natural phenomena – why volcanoes erupt, why the seasons change – humans looked to God for answers. In shaping the social world, from economics to love, religion served as our guide.
Humans, Kant argued, always had the capacity for reason. They just hadn’t always had the confidence to use it. But with the American and later the French Revolution, a new era was dawning: reason would replace faith, and the human mind, unshackled from authority, would become the engine of progress and a more moral world. “Sapere aude!” or “Have courage to use your own understanding!”, Kant urged his contemporaries.
Two and a half centuries later, one may wonder whether we are quietly slipping back into immaturity. An app telling us which road to take is one thing. But artificial intelligence threatens to become our new “other” – a silent authority that guides our thoughts and actions. We are in danger of ceding the hard-won courage to think for ourselves – and this time, not to gods or kings, but to code.
ChatGPT was launched only three years ago, and already one global survey, published in April, found that 82% of respondents had used AI in the previous six months. Whether deciding to end a relationship or who to vote for, people are turning to machines for advice. According to OpenAI, 73% of user prompts concern non work-related topics. Even more intriguing than our dependence on AI’s judgment in daily life is what happens when we let it speak for us. Writing is now among the most common uses of ChatGPT, second only to practical requests such as DIY or cooking advice. The American author Joan Didion once said: “I write entirely to find out what I am thinking.” What happens when we stop writing? Do we stop finding out?
Worryingly, some evidence suggests that the answer might be yes. A study by the Massachusetts Institute of Technology used electroencephalography (EEG) to monitor the brain activity of essay writers given access to AI, search engines like Google, or nothing at all. Those who could rely on AI showed the lowest cognitive activity and struggled to accurately quote their work. Perhaps most concerning was that over a couple of months, participants in the AI group became increasingly lazy, copying entire blocks of text in their essays.
The study is small and imperfect, but Kant would have recognised the pattern. “Laziness and cowardice,” he wrote, “are the reasons why so great a proportion of men … remain in lifelong immaturity, and why it is so easy for others to establish themselves as their guardians. It is so easy to be immature.”
Sure, AI’s appeal lies in its convenience. It saves time, spares effort and – crucially – offers a new way to offload responsibility. In his 1941 book, Escape from Freedom, the German psychoanalyst Erich Fromm argued that the rise of fascism could be explained in part by people preferring to surrender their freedom in exchange for the reassuring certainty of subordination. AI offers a new way of surrendering that burden of having to think and decide for yourself.
AI’s greatest allure is that it can do things our minds can’t – sift through oceans of data and process it at unprecedented speed. Sitting in the car in Marseille, this was, after all, why I chose to trust the machine instead of my friend in the passenger seat (a decision she took as an insult). With access to all the data, surely the app must know best – or so I thought.
The problem is that AI is a black box. It produces knowledge, but without necessarily deepening human understanding. We don’t really know how AI reaches its conclusions – even the programmers admit as much. Nor can we verify its reasoning against clear, objective criteria. So when we follow AI’s advice, we are not guided by reason. We are back in the realm of faith. In dubio pro machina: when in doubt, trust the machine – that may become our future guiding principle.
AI can be a formidable ally to humans in rational inquiry. It can help us invent drugs, or free us from “bullshit jobs”, or doing our taxes – tasks that demand little thought and offer little satisfaction. All the better. But Kant and his contemporaries did not plead the case of reason over faith just so humans could build better shelves or have more spare time. Critical thinking was not just about efficiency – it was a practice of freedom and human emancipation.
Human thinking is messy and full of errors, but it forces us to debate, to doubt, to test ideas against one another – and to recognise the limits of our own understanding. It builds confidence, both individually and collectively. For Kant, the exercise of reason was never just about knowledge; it was about enabling people to become agents of their own lives, and resist domination. It was about building a moral community grounded in the shared principle of reason and debate, rather than blind belief.
With all the benefits AI brings, the challenge is this: how can we harness its promise of superhuman intelligence without eroding human reasoning, the cornerstone of the Enlightenment and of liberal democracy itself? That may be one of the defining questions of the 21st century. It is one we would do well not to delegate to the machine.
