Back to Blog
📖 Tool Tutorials 管理员 · · 3 minutes · 239 Views

A Chinese Literature Professor Said AI Can't Write Good Essays, So I Sent Him a Prose Piece

This article challenges a Chinese literature professor's view through a hands-on experiment: using an AI writing assistant to compose a prose piece. AI can produce polished text, but it fails to capture real-life details and emotional warmth. The conclusion: AI is a helpful tool, but the 'soul' of a good essay comes from human experience and feeling.

A few days ago, I saw a post on my social feed from a Chinese literature professor saying that AI can't write good essays, and that so-called 'AI-generated creative writing' is at best a patchwork of words—no flesh, no blood, let alone any soul. The comment section was in an uproar; some said the professor was too stubborn, others thought AI would eventually fail. I wasn't going to bother getting involved, but then I thought, isn't this a perfect opportunity? I happen to have an 'AI writing assistant' that I usually use for work reports and promotional copy. Today, I decided to test it and see if it could rise to the challenge.

I didn't choose a grand theme; I wrote a prose piece titled 'The Wonton Soup at the End of the Alley.' It was about the old street across from my childhood home, on a winter evening, with dim streetlights, and the old man selling wontons lifting the pot lid, steam rushing into his face, mixed with the aroma of lard and scallions. I fed all the details I remembered to the AI: the wet mud in the cracks of the blue stone slabs, the calluses on the old man's hands, the chipped edge of the bowl, the seaweed and dried shrimp floating on the soup surface. The AI quickly returned a draft. To be honest, at first glance, the sentences were smooth, the imagery was there, and it even managed a sentimental line like 'that hot soup made my eyes sting.' I raised an eyebrow—okay, not bad.

But the more I read, the more something felt off. It wrote 'the wind at the alley's mouth carried whispers of time' and 'wontons are the gentlest footnote of time.' Pretty, yes, but all I could think of was the flavor of 'standard model essays' from Chinese textbooks. It described the wontons beautifully, but it couldn't capture the old man's fingers whitened by coins when collecting payment, or my silly self squatting on the curb waiting for my bowl, nearly falling over when the neighbor's dog barked at me. It wrote 'the hot soup made my eyes sting,' but it didn't capture that the sting was also laced with longing for my grandmother—that bowl of wontons was the first I ate alone after she passed away.

I screenshotted the AI's draft and sent it to the professor with a note: 'Take a look, this is AI-written. Do you think it's a good essay?' He replied quickly: 'The writing is neat, the technique is polished, but there's nothing memorable. A good essay makes you want to flip back to a certain line after you close the page, and then stare blankly for a while. Can you stare blankly at this one?' I stared at the screen and suddenly felt he was right. AI can mimic grammar, pile up imagery, and even guess emotional buttons like 'warmth' and 'nostalgia,' but it doesn't understand why a person might suddenly want to cry on a winter night, or what that bowl of wontons meant to that child—after losing his grandmother, it was the first time someone handed him something warm.

Of course, I'm not completely dismissing AI. That day, I asked it to help revise the ending, and it gave me a version: 'Later, that street was demolished, but that warmth still lingers.' That was a nice touch. But the backbone of the essay—those moments I swallowed, laughed at, and secretly wiped tears over—AI can never write those. It's like a clever apprentice who can grind ink, lay out paper, and polish the brush for you, but the tremor when the brush touches the paper has to come from you.

Later, I turned this experience into an article and published it. The professor liked it and commented: 'This one, AI couldn't write.' I replied: 'Right, but it helped me bring the bowl of wontons.' This made me think for a long time. Maybe AI isn't here to take our jobs; it's more like a tool that can chat with you and help you build a framework, but the real flavor has to come from what you put into the pot yourself. Like that bowl of wontons—the broth was AI's doing, but the dried shrimp and seaweed, I sprinkled those myself.

239 Views · 3 minutes

🔗 Related Tools

Try these practical tools related to this article

📝 Related Posts

You might also like these articles

tool-tutorials

Beware! Your IP Address Is Exposing Your Home Address

An IP address is not just a meaningless string of numbers; it's like your home's network house number. With lookup tools, it can directly locate your city, street, or even residential complex. Everyday activities like browsing, commenting, or connecting to Wi-Fi can leak your IP, which, if exploited, can lead to harassment or even fraud. This article reminds everyone not to casually fill in personal information, to be cautious with public networks, to change default device passwords, and suggests checking your own IP to see what's exposed, raising awareness to protect privacy.

09-09
tool-tutorials

The days of not understanding JSON formatting and being blamed by colleagues are over

This article explains in plain language the practical value of JSON formatting tools. Starting from daily scenarios like colleagues shifting blame, taking over messy projects, and integration testing comparisons, it shows that formatting tools can help you clarify data, quickly report errors, and precisely compare differences, saving time and effort while avoiding blame. It emphasizes that knowing how to use tools is true intelligence, teaching you to solve the most annoying problems with minimal cost, and saying goodbye to the hard days of searching through messy JSON.

09-09
tool-tutorials

Changed a JS compression config, online errors dropped by 80%

Frequent JS errors online? Don't rush to blame the logic; 80% of the time it's the compression config causing trouble. This article uses real cases to show you how the compressor's default "smart" settings can ruin old projects, from mis-deleting conditional branches to messing up variable names. Step-by-step, I'll teach you to turn off a few key optimization options, cutting error rates by 80% while only increasing code size by 15%, in exchange for stable operation—totally worth it.

09-09