By Jane Sonnenschein · March 19, 2026

Today, while scrolling through Xiaohongshu, I came across a creator talking about changes to WeChat Official Accounts—a publishing system built into WeChat that is somewhat comparable to Medium, though it is also closely tied to subscriptions, messaging, and the wider WeChat ecosystem. Something she said stayed with me.

Her father had asked her the day before, “Why do all the writers I follow seem to have stopped posting? Why don’t I see them anymore?”

She looked into it and realised that they had not stopped writing at all. The way WeChat distributes articles has simply changed. It is no longer the straightforward system it once was, where you followed someone and then saw their latest posts in chronological order. Now, unless you mark an account as a favourite or deliberately open certain less visible sections, you may not see updates from people you have already chosen to follow.

As she explained this, she sighed and said something along the lines of:

WeChat Official Accounts used to feel like one of the last relatively clear streams among all these platforms. Now even that stream seems to have been pulled into the same current.

When I heard her say it, I felt a strange sense of loss.

It was difficult to describe. It was not simply that a feature had disappeared or a menu had changed. It was the sudden realisation that even the basic right to say, “I followed this person, so I should be able to see what they publish,” was slowly being taken away by the platform.

And this is no longer just a WeChat problem.

Xiaohongshu, Douyin, Taobao, and many of the other platforms we use every day are all moving in roughly the same direction. Platforms are becoming less interested in letting you decide what you want to see. Instead, they increasingly want to decide what you should see next.

The internet did not always work this way.

When we used search engines, there was already far too much information, but at least the search still began with us. We typed in a question, looked through the results, compared different sources, and tried to decide for ourselves: Is this really what I am looking for? Is this claim reliable? Why is this result ranked above the others?

We might still have reached the wrong conclusion, but the acts of searching and thinking were at least our own.

This is one reason my husband has always been deeply cautious about AI.

When he was writing a project paper for an IT internship, I once suggested that he could use AI to help him work through the parts where he felt stuck. He tried it, but quickly gave up. His opinion was very clear: it was not right.

He did not mean that the sentences were unreadable or that the result was complete nonsense. The problem was that AI could not reach the genuinely important and professional point he was trying to make. Especially when he was writing in German about IT, he felt that much of the generated material was immediately recognisable as imprecise, shallow, or simply not what he intended to say.

So he preferred to search on Google himself.

For him, the value of a search engine was not merely that it gave him an answer. It allowed him to open official websites, follow links to original materials, look for sources he considered reliable, and then decide which information was useful and which claims could actually stand up.

In other words, it was not that he disliked convenience. He simply understood that once you outsource the act of judgement as well, you may save time on the surface while losing control over the information itself.

Yet this is exactly the step that today’s platforms increasingly want to take away from us.

You no longer need to search. You only need to keep scrolling. You do not even need to know what you want to see, because the platform will decide faster than you can. It learns what you click, what you watch, where you pause, what you like, what makes you anxious, and what keeps you hooked. Then it places one thing after another in front of you, choosing whatever it believes is most likely to keep you there.

On the surface, this looks like the platform understands you better.

Seen from another angle, however, it is gradually weakening a particular set of abilities: the ability to search actively, to filter information, and to judge things for yourself.

That is what truly makes me uncomfortable.

When people spend long periods in a state of being constantly “fed” information, they may slowly lose the desire to search for anything themselves.

This does not happen because people suddenly become less intelligent. It happens because the path designed by the platform is easier, faster, and more immediately rewarding than active thought.

Short videos work in the same way.

You may begin by telling yourself that you only want to relax for a few minutes. Then one video follows another, and suddenly the entire afternoon has disappeared. While you are scrolling, it feels exciting, as though you are constantly taking in something new. But when you stop, a strange emptiness and panic appear: Where did all that time go? What did I actually watch? It feels as though I saw everything, yet retained almost nothing.

Many people know this feeling now.

A friend of mine is a particularly striking example.

I have always thought of her as an exceptionally disciplined person. She gets up early to exercise every day and has done so for many years. She pays close attention to what she eats, and she has kept practising Chinese calligraphy. She may not necessarily be the kind of person people call naturally gifted, but her discipline is remarkable. Looking at her, you would assume that her life had a steady rhythm and that she would not easily be pulled around by the outside world.

Yet even she once told me something that shocked me.

She said that when she had nothing planned in the afternoon, she would sometimes sit down and start scrolling through Douyin. At first, she only meant to watch for a little while. Before she knew it, the whole afternoon was gone. What made it worse was that she felt bad afterwards. She had the feeling that she had not really taken anything in, yet the time had already passed. And when she did not scroll, she felt slightly uncomfortable.

When I heard this, I felt a sudden chill.

It showed me that the most frightening thing about short videos is no longer simply that they make people lazy. Their real power lies in the fact that even someone who is already extremely disciplined may struggle to resist the mechanism that keeps pulling her downward.

And the problem no longer ends with platforms choosing content for us.

Since the rise of AI, many people have started outsourcing not only the search for information, but also the acts of thinking and judging.

The more I learn about AI, the more strongly I feel this.

In class today, our teacher asked us to use an overseas platform to generate music. I had worked more with videos and dialogue before, but this was my first time using AI to create music in a more systematic way. The experience was astonishing. Lyrics appeared in seconds. A melody followed almost immediately. The efficiency was extraordinary, and the music itself was actually quite good.

Seen against several thousand years of human history, this kind of convenience is almost unprecedented.

It is difficult not to be impressed.

In the past, writing lyrics or composing music meant struggling, revising, getting stuck, and starting again. Inspiration might arrive and disappear before you could catch it. When inspiration dried up, that was painful too. Creation has never been an easy process.

AI does not go through any of that.

It does not hesitate. It does not struggle. It does not experience the difficulty of asking, “How do I really want to express this?” It processes data, rearranges patterns, and produces a result.

And that is precisely where the problem begins.

All the slowness, difficulty, hesitation, and repeated self-correction that people once experienced may seem inefficient. Yet those moments were often exactly where real thinking took place.

Now, more and more people are not simply using AI to improve efficiency. They are slowly handing over the entire process:

I no longer need to search. AI will search for me.

I no longer need to read the original source. AI will summarise it.

I no longer need to decide whether something is true. AI will give me the answer.

I do not even need to work out what I want to say. AI will write it for me.

There is nothing inherently wrong with using AI for support from time to time. But when someone does this continuously, what they may eventually lose is not just one particular skill. They may lose an entire way of using their own mind.

Over the past few days, I also came across a very typical example of what people call “AI poisoning”.

Someone invented a fitness tracker that did not exist, then produced large amounts of apparently professional reviews, promotional articles, and product descriptions and published them online. Some AI systems later treated this fictional product as real. They even confidently recommended it to users.

Many people would say that this simply proves AI can make things up.

But the more I thought about it, the more I felt that the most frightening part was not merely that the AI gave a wrong answer. It was that more and more people may stop asking why the answer is wrong, or why they should not immediately believe it.

For many users, AI is no longer simply a tool. It is increasingly treated as something that has already done the thinking on their behalf.

That is what I find dangerous.

In the past, even when someone was being lazy, they usually knew they were being lazy.

The frightening thing about platforms and AI is that they may slowly remove even that awareness. You feel as though you are constantly receiving information, constantly learning, and constantly finding answers more efficiently. Yet your ability to search, judge, and build meaning for yourself may be weakening little by little.

Platforms are not afraid of this.

On the contrary, they may even prefer it.

A person who has grown used to being pushed, fed, summarised for, and thought for is the most stable kind of user. Such a person requires little effort to keep engaged and is far more likely to remain on the platform.

This is why I increasingly feel that the most frightening thing about platforms is not simply that they make us addicted or waste our time. It is that they quietly change the relationship between people and information.

They encourage us to search less, choose less, and think less. Eventually, they may even persuade us to outsource judgement itself.

And that was when I suddenly thought of the film Idiocracy.

It is an older black comedy about a completely ordinary man, someone who is not particularly successful in the real world, who accidentally falls asleep for a very long time. When he wakes up, he finds himself five hundred years in the future.

But the people of the future have not become more intelligent or advanced. Instead, the judgement and intelligence of society have declined to such an absurd level that this once unremarkable man is suddenly considered highly intelligent.

When I first watched the film, I mainly saw it as exaggerated satire.

Today, however, I suddenly understood that the reason it makes people laugh is precisely because the thing it touches is not funny at all.

Of course, we are still nowhere near the extreme world shown in the film.

But some of the tendencies are already here.

Platforms are becoming better at choosing for us.

AI is becoming better at answering for us.

Machines are becoming better at speaking.

And if people are not careful, they may gradually lose the desire to search, think, and judge for themselves.

The irony is that many platforms are now trying to crack down on mass-produced AI content while simultaneously pushing for faster content production, smoother recommendations, and more precise ways of feeding people whatever will hold their attention.

The result is a particularly absurd situation.

Platforms are becoming busier and more crowded.

Yet the feeling of encountering a real person is becoming harder to find.

Everywhere, there are smooth sentences, rapid opinions, polished structures, standardised emotions, and efficiently produced content.

But after reading enough of it, you begin to feel empty.

What truly moves people has never been information alone. It is the trace left behind when a person has lived through something, struggled with it, hesitated, thought about it, and then finally put it into words.

That is why I am increasingly tired of content that immediately feels as though it was written by a machine.

It is not necessarily because the content is wrong. It is because there is no person inside it.

Writing that still feels human is not always more perfect, faster, or smoother.

It may be slower, more winding, and even a little clumsy.

But it carries the marks of a life that has actually been lived. It contains someone’s judgement, someone’s hesitation, and the process of getting stuck and then slowly finding a way forward.

These are precisely the things that platforms and AI find hardest to possess.

So why did a video about changes to WeChat Official Accounts lead me all the way here today?

Because I increasingly feel that the real challenge facing our generation may no longer be simply whether AI will replace us.

The greater challenge may be this:

When platforms become better at choosing for you and AI becomes better at answering for you, can you still hold on to your ability to search, judge, and think for yourself?

If you cannot, then the problem is no longer merely about efficiency.

People themselves will begin to change.

And by then, the most frightening thing may not be that machines are becoming more like humans.

It may be that humans are slowly becoming more like machines.

This essay is also available in other languages:

Chinese version: 从公众号改版到AI投喂：我为什么突然想起了《蠢蛋进化论》

German version: Von der Umstellung der WeChat Official Accounts bis zum Füttern durch KI: Warum ich plötzlich an Idiocracy denken musste