Lately social media is full of similar boasts. “I sent ChatGPT a selfie and it told me my personal color season for free.” The diagnostic image on the screen looks almost indistinguishable from the results of an in-person consultation. The verdict that follows is always the same: a diagnosis service that used to charge around $80 is destined to disappear. AI interior simulators, AI hair simulators, AI fashion coordinators all stand in the same line.
The question that usually follows is, “So how do we survive?” Data moats, regulatory walls, domain expertise — the same set of advice cycles through. But the more fundamental question lies elsewhere. What were we actually buying with that $80 in the first place?
Think about it. People weren’t paying $80 for a single line of color code. The voice that said “this color looks really good on you” every time you held a swatch under the standardized lighting, the half-hour of someone seeing only you and no one else, the credential of being able to tell a friend at a café “I got diagnosed as Warm Spring” — those things were all bundled into that price. The color may have been the smallest part of the bundle.
AI broke the bundle apart. By pulling out only the functional piece — the color diagnosis itself — and giving it away for free, AI revealed that we hadn’t actually been buying color. The recording industry went through something similar first. Back when a $15 CD was the unit of sale, what people bought wasn’t a single song but a package of jacket art, lyric booklet, and the satisfaction of owning a physical object. When streaming separated the package into songs at ten cents apiece, what we had actually been buying finally became visible. AI is doing the same thing. What it killed was not the business itself, but the era in which “attention paid to me” could carry a price tag.
The same thing is happening everywhere. Half of the $250-per-hour lawyer consultation wasn’t legal research; it was the acknowledgement, “your situation truly is unfair.” A large share of a gym’s personal-training fee wasn’t the exercise prescription; it was the attendance-check of “you came in today.” Nutrition counseling, learning coaching, even parts of a doctor’s visit — all the same. When AI makes the informational layer free, the attention, the recognition, the care that was packaged with it stands exposed.
So the polarization that’s coming is not simply a polarization of price. It is a polarization of meaning. Information and diagnosis converge toward free; in their place, the human gaze, time, and recognition that used to sit inside them become more expensive instead. Where the entry-level consultation disappears, a premium service that says “I will spend two hours on no one but you, the day before your wedding” remains. Businesses didn’t fail. They’re catching up, belatedly, to what they were actually selling. Korean society is likely to feel this shift especially fast. The more accustomed a market is to purchasing external authority to define oneself within a short window of time, the bigger the shock when that authority turns free.
The truly dangerous person is the one who, never having figured out what they were selling, complains that “AI killed my business.” What AI took was not the color diagnosis. It was the time you had to look honestly at what value your work was actually delivering. Avoid that question, and no moat will save you.
The same goes for a working professional’s career. The salary your company pays you — what is it exactly compensating? Report writing? Information organization? Running meetings? All of these will soon be near-free thanks to AI. Then what was the company really buying with your salary? The willingness to take responsibility, the trust your colleagues have in you, the judgment you bring to hard decisions. If it’s any of those, you’ll be more expensive in the AI age, not less. On the other hand, if you’ve defined your work as nothing more than “information processing,” that’s exactly the definition by which AI will replace you.
What AI killed is not the business. What AI killed is the era when you didn’t have to ask what your work really was.