As artificial intelligence settles into everyday life, we keep reaching for metaphors to grasp its essence: a tireless assistant that never sleeps, a parasite that devours our dignity, an oracle with a habit of guessing and apologizing. Each metaphor does its own work. The assistant flatters us; the oracle mystifies us. But metaphors are never merely figures of speech. The one a society embraces quietly shapes what we invest in, what we fear and what we choose to regulate.
The metaphor I want to argue for is: AI is a mirror of who we are. The algorithms behind these systems are designed to reproduce aspects of human thought, while the vast quantities of human writing and behavioral data on which they are trained carry the imprint of our reasoning, our prejudices and our desires.
Ideally, a machine of this kind would help us sort out information and draw reasonable inferences from the evidence. Yet human intelligence has never confined itself to such boundaries, and a mirror does not choose what to reflect. If the reflection disturbs us, we should be careful about where we direct our complaints.
Scholars continue to warn that AI could threaten human safety and even human existence. I do not dismiss these dangers. But their roots lie not only in the machine's capabilities; they also lie in human appetites: developers racing for a first-mover advantage, executives pursuing profit, and users eager to adopt the next new thing. Three examples illustrate how quickly AI's supposed neutrality collapses when human interests get in the way.
Consider the robots that box. In July, a Chinese robotics company launched what it billed as the world's first humanoid fighting league. In one bout, a robot reportedly continued swinging after its head came off. The machine itself bears no grudge. It follows its programming, faces an opponent, and is expected to win. The violence on display is therefore ours: staged and sold for entertainment. Yet coverage can celebrate the spectacle and, almost in the same breath, voice concern about safety, as though the entertainment and the danger belonged to separate worlds.
Weapons, drugs and other dangerous technologies have taught us a lesson we should not have to learn again: safeguards must be established before disaster strikes, not after. That conversation around AI remains inadequate, partly because public understanding of the technology is thin and partly because calls for oversight are too easily dismissed as hostility to innovation.
Consider, next, the banks. One Korean bank recently disclosed that customers' personal information, including names, addresses, phone numbers and other identifying details, was exposed in an incident attributed to malicious AI-assisted hacking. Subsequent reports suggested that the impact extended to seven institutions and tens of thousands of people.
What strikes me is what often goes missing from the coverage. Headlines marvel at AI's capacity to hack, turning a security breach into another story about the threat posed by AI itself. Far less attention goes to the people behind the attack, their motives or the defenses that should have been in place against a hardly unforeseeable threat.
Had a criminal gang stolen those records, we would be discussing suspects, prosecutions and prison sentences. When an algorithm becomes part of the operation, it may change the speed, scale or cost of the crime. It does not erase the human agency behind it.
Fraud follows the same pattern. Forged official documents, manipulated photographs, fabricated videos: These are established forms of deception that long predate generative AI. People have always lied, distorted the truth and blurred the boundary between fact and fiction. Deception is one of our oldest habits, and the industries built around exploiting it are almost as old.
Why, then, should we be surprised, let alone terrified, when that habit acquires a more powerful tool? There is, in one sense, something reassuring in the fact that AI reproduces human deception and scheming alongside our other characteristics. The irony is worth considering: The machine did not invent malice. It learned from ours.
A familiar Korean storytelling motif captures the rest. A man walks home at night, sees leaves trembling in the darkness, and imagines ghosts. The next day, in a cheerful mood, he watches the sea toss those same leaves about and finds them beautiful. The scene changes little; the attitude changes everything. We might call this the complexity of human feeling. Less flattering terms would be fickleness and psychological fragility.
The machine, poor thing, has neither our humanity nor, on present evidence, consciousness of its own. Yet we began personifying it almost as soon as it arrived. Now some would have us regard it as a god.
So let me address my fellow citizens who fear AI. Much of our anxiety, I suspect, comes from a limited understanding of our own capacity for harm and from our reluctance to confront what history has already taught us about ourselves.
I do not own a crystal ball, though. Some report is bound to hit the wires about AI's demonry, and the cycle of shivers will start again. How should we respond? The answers remain open to debate, but I hold one conviction firmly: A look back at human history makes one thing painfully clear. We have already committed atrocities on a scale that no AI has ever approached.
If that conviction leads us anywhere, it should lead us here. The central question is not simply whether AI will behave badly, but who is building it, whose money is financing it, and what those investors and builders intend to do with it. A society that scrutinizes the machine while waving through its owners has put its magnifying glass over the wrong half of the picture.
Watch the mirror, by all means. But watch more closely the people behind AI — you and I included: the engineers who design these systems, the funders and CEOs who determine which capabilities are pursued and how the technology is deployed. Hold them accountable. The reflection will improve only when its source does.
Lim Woong
Lim Woong is a professor at the Graduate School of Education at Yonsei University in Seoul. The views expressed here are the writer’s own. — Ed.
khnews@heraldcorp.com
