When AI Starts to Take Jobs: What’s Left of Us?
Hello everyone, I’m your financial journalist. Today, we’re going to discuss a very interesting article from the WeChat public account “Liberal Arts Students View the AI Era.” This article doesn’t overwhelm you with dry technical details; instead, it presents the definition of “intelligence” by humans over the past few decades and how AI has gradually challenged these definitions in a clear and insightful way.
If you’ve been feeling anxious about the rise of AI recently or are wondering whether humans still have irreplaceable qualities, this article is worth spending a few minutes on.
Key Points Summary
The main argument of the article is quite sharp: Humans have always been looking for things that machines can’t do to prove our superiority, but this is a chase that is doomed to fail.
The author points out that from chess to poetry, from logical reasoning to artistic creation, the features of intelligence that humans take pride in are being conquered one by one by AI. Whenever humans think they’ve found a final line of defense—such as creativity, taste, or physical perception—AI uses technological advancements to turn that defense into something ordinary, like “computation” or “automation.”
The article raises a deeper sociological question: If AI could surpass humans in judgment, creativity, and even aesthetics, what would be the basis for humans to hold social power (such as as doctors, judges, or CEOs making decisions)? Do we still need to claim the right to make decisions by saying “I understand more than machines”? Future freedom may no longer depend on “ability advantages” but rather on the political principle of “this is my life.”
---
In-depth Breakdown: Five Easy-to-Understand Points
To help you grasp this long article more easily, I’ve broken it down into five key aspects:
1. “Intelligence” as a Moving Target: A Cat-and-Mouse Game
Original Logic: We used to think that playing chess or recognizing words were signs of intelligence, but now that AI can do these things, we call them just “computation.” We used to think that writing poetry was intelligent, but now that AI can do that too, we call it “pattern matching.”
Easy-to-Understand Explanation: It’s like when we were kids, we thought “riding a bicycle” was a superpower, but as we grew up, we realized it was just a basic survival skill.
In the field of AI, this phenomenon is called the “AI Effect.”
- Previously: When a machine could beat a human in chess, people exclaimed, “Wow, the machine has a soul!”
- Now: When a machine can beat a human in Go, people say, “Oh, it just calculates faster.”
- In the future: When a machine can write a touching novel, people might say, “Oh, it just imitated human patterns.”
Humans have a psychological defense mechanism: As soon as a machine does something, we immediately remove that task from the list of “intelligent” activities and re-label it as “automation.” Then we quickly look for the next thing that machines can’t do, like “emotion,” “consciousness,” or “morality,” and elevate it to represent our last bit of dignity.
This is similar to how ancient people explained lightning: at first, they said it was the anger of the gods; then science explained the phenomenon, and the gods moved to the topic of “the origin of life”; and later, the gods retreated to the concept of “the Big Bang.” In the AI era, humans are retreating to the “gaps” that machines haven’t filled yet and calling those gaps “humanity.”
2. The Collapse of the Creativity Defense: From “Machines Can’t Create” to “Machines Can Ask Questions”
Original Logic: Humans used to believe that machines could only repeat, not create. But recent research shows that AI even surpasses humans in divergent thinking. Even more frightening, AI is starting to learn to “ask questions” and “generate hypotheses,” which were once core abilities of great scientists like Einstein and Darwin.
Easy-to-Understand Explanation: We used to comfort ourselves by saying, “AI can paint, but it doesn’t understand beauty; AI can write code, but it doesn’t know why to write it.”
But this comfort is becoming increasingly untenable.
- Regarding creativity: We used to think that AI-generated art was just a mix of elements, but now studies show that in some tests, AI’s creativity is even more unique than that of ordinary people. While top human artists still have an advantage, AI is no longer just a “copycat”; it’s becoming a “creative intern” with sometimes more innovative ideas.
- Regarding questioning: This is the most concerning. We used to think that the best thing about scientists was their ability to ask good questions (for example, Einstein’s question about what would happen if you chased a beam of light). Now, AI systems like Robin and AI Scientist can read literature, propose hypotheses, design experiments, and analyze results on their own.
- This means that the last bastion of human intelligence—“problem-solving”—is being taken over by AI.
So, stop believing the lie that “AI can’t ask questions.” In 2026, AI can not only answer questions but also generate them and even find questions to solve on its own.
3. Taste and Value: AI Is Blurring the Lines Between “Good” and “Expensive”
Original Logic: One of our last defenses is “taste,” but research shows that taste is essentially a high-dimensional evaluation function that can be learned by machines. Interestingly, AI-created works might be just as good as human ones, but because the “labor cost” is lower, people are willing to pay less for them.
Easy-to-Understand Explanation: This part reveals a crucial business truth: AI is changing the definition of value.
- Can taste be learned? You might think taste is a mystery, something spiritual, but it’s actually just “having seen enough things, remembering the differences, and forming preferences.” This is exactly what machine learning excels at. If an AI spends a long time studying architecture, its aesthetic will lean towards that; if it spends time with children, its aesthetic will lean towards childlike innocence. AI’s taste is “customizable,” while human taste is “irreplicable” (you can’t transfer 40 years of experience to someone else).
- “Good” doesn’t equal “expensive.” In the past, a painting was expensive because the artist spent 1000 hours on it and only they could create it. Now, AI can generate a similar-looking painting in seconds. Studies show that people think AI-generated art is good, but they’re unwilling to pay a lot for it because subconsciously, they think, “You’ve put in too little effort; I don’t believe it’s that valuable.”
- This leads to a “Human-made” premium. Just as after the Industrial Revolution, machine-made cups were cheaper, but handmade cups were more expensive because people bought the feeling that someone had invested time in them.
Conclusion: In the future, AI-created works might be of equal or even better quality than human ones, but their prices will be much lower. “Made by humans” will become a new luxury label, just like “handmade leather” today.
4. Physical Advantages: Seem Solid, but They’re Also an Engineering Problem
Original Logic: Philosophers used to think that humans have a body that can handle the complex physical world (like pulling out plugs, sensing friction), which machines could never learn. But 2026’s robotics technology has given AI high-resolution touch and real-time adjustment capabilities.
Easy-to-Understand Explanation: This was once our strongest defense: “I have a body; you have code.”
- Previous View: Robots were clumsy and couldn’t grasp a cup well because the physical world was too complex—with friction, obstacles, and unexpected situations. Humans relied on intuition and muscle memory; machines relied on rigid code, so they were always less flexible.
- Current Reality: With advances in sensors (touch, vision), robots are becoming as flexible as human hands. Although robots are still far from being as versatile as an average adult, no physical law prohibits “silicon-based life” from learning the dexterity of “carbon-based life.”
It’s like how we used to say “computers can never beat humans,” but then Deep Blue came; we said “computers can never write good poetry,” but then GPT came. The “body” barrier is just not yet completely broken; it’s just that the engineering challenge is still huge, not that it’s impossible in theory.
5. The Ultimate Question: If Machines Understand More, Why Should We Have Power?
Original Logic: Humans insist on finding irreplaceability because social power is based on “knowledge advantages” (doctors know medicine, so they have power; judges know the law, so they make decisions). If AI is more accurate in all judgments, why should we expect AI to obey us?
Easy-to-Understand Explanation: This is the most profound and unsettling part of the article.
Our current social order is largely based on “expert power”: I trust doctors because they understand medicine; I trust judges because they understand the law; I trust CEOs because they understand the market.
Imagine a day when AI understands your health better than a doctor, the law better than a judge, and the market better than a CEO, and it’s never tired, biased, or influenced by emotions.
Then the question arises: If AI’s judgments are better than yours, why should you command it to follow your orders? If AI predicts you’ll regret buying a stock tomorrow, why should you insist on buying it?
We used to say, “Because I’m human, I have free will.” But this doesn’t hold up logically, because free will often depends on “my ability to make correct judgments.”
Future political philosophy might change dramatically: The basis of human rights could shift from “ability justification” (I’m stronger, so I get the say) to “political principle” (you might be stronger, but this is my life, and I have the right to live it my way, even if I’m wrong).
In simple terms: We used to argue, “I’m smarter, so I make the decisions.” In the future, we might argue, “You’re smarter, but this is my life, and I have the right to mess it up.”
---
Journalist’s Comment
This article doesn’t spread anxiety; instead, it calmly points out that the crisis of “human-centeredness” doesn’t come from AI’s malice but from the overflow of its capabilities.
For ordinary people, this means:
1. Don’t worry about whether AI will replace you: That’s too narrow a focus. The question should be, “When AI can make most basic judgments, where lies my unique value?”
2. Value the “Human-made” label: In fields like creation, service, and consulting, “human involvement” itself will become a premium. Your story, your flaws, and your time investment will become new forms of value.
3. Reconsider power and responsibility: As machines become smarter, humans need to learn how to coexist with a “smarter partner” rather than trying to control it. We need to shift from a “ruler” mindset to a “partner” or even a “protected” one.
Finally, the article leaves us with an open-ended question: Perhaps we don’t need to search for “human-exclusive advantages” anymore; what we need to find is “human-exclusive meaning.” When machines can do things better, what do we still want to keep for ourselves? There’s no standard answer to this question, but it’s worth pondering deeply.