Abstract:
Not long ago, Claude, one of the most mainstream AI tools at present, underwent a wave of updates. One of the updates was to optimize the Chinese user experience. Claude added Simplified and Traditional Chinese UI interfaces to both the desktop and web pages. China was added to the country and region options, and UnionPay payment was also supported.
As one of the best AI tools for long text reading, Claude has a very large Chinese user base. According to past practice, this kind of comprehensive Chinese-oriented optimization will undoubtedly greatly improve the experience of Chinese users. Although Claude is not cheap, on the surface, this is "universal celebration" news.
However, after Claude updated the Chinese language, everyone’s immediate reaction was not happiness——
But hesitant, defensive, and scolding.
It’s not that everyone doesn’t welcome a tool that is very easy to use to update Chinese, but everyone is very worried that this is another “fishing law enforcement” by Claude.
Today, as AI has gradually become a part of more and more people’s daily lives, it goes without saying that they have also had their own incarnations in popular culture, such as Doubao’s noble and cool goddess memes, and DeepSeek’s gluttonous fat fish.
Many AI short dramas featuring these AI anthropomorphic images have also appeared on the Internet, depicting the "Friends"-style daily life of different AIs renting houses and working together, which can be called a brand new landscape in the AI era.
Picture source: Bilibili@东冰
For those who are not very familiar with various AI tools, the most famous and out-of-circle aspect of Claude, the AI, is that it almost catches and bans Chinese people in a "hunting" manner.
There is a meme that has been particularly popular in the past two days. It is the account owner playing a bottom line testing game with Claude, which is to let the AI constantly ask itself questions to test the bottom line. This kind of game is not uncommon in the interaction between netizens and AI. Just like in the AI dialogue, people try every means to get AI to help them write a "human extinction plan". Essentially, it is not to test moral standards, but to see what outrageous remarks the AI will make in those particularly difficult problems.
The "test results" of this account owner were also unbelievable - however, the reason that went against the sky was not that Claude said something outrageous, but that Claude came up and asked if the account owner was Chinese. After getting the affirmative answer, he blocked him. It was clean, ruthless and ruthless, just like a short novel ten years ago, with an O. Henry-style ending.
And this kind of "thriller comedy" can be said to be more and more common after the UI was recently updated to Chinese. As we all know, even if AI is used for work, many people will inevitably regard AI as the only person to talk to when they have nowhere to talk. Many people even regard AI as their most important friend because of their frequent heart-to-heart conversations. This is not difficult to understand. It is difficult for anyone to respond to everything you want to say 24 hours a day like AI.
In Claude, this kind of conversation between family and friends also exists. But the risk is that if you are a Chinese user, after a thoughtful and warm response, it may stab you every minute. It can be called a new generation of minesweeper dating. AI and humans form a new SM relationship——
Honestly talking to friends is boring.
Catching Chinese people and banning them is actually not news to Claude. Claude has always been famous for being quick and accurate when banning people. Whether you are here to chat or work, he will ban you as soon as you are told and never delay. This kind of ban is implemented equally and will not be affected by your payment behavior at all. Even if you are a paying user or even a premium paying user, they will still block it without refund.
According to the official announcement of Anthropic, the team behind Claude, 11.4 million accounts were banned by them in the first half of this year. Nearly 400,000 of them chose to appeal, but in the end, only more than 40,000 were unblocked. Claude has therefore become the most famous cold-faced ban king in the AI arena.
Among these banned users, the ones who howled the most and were banned the most were Chinese users.
A clear background here is: From a legal point of view, there is no problem for Claude to ban Chinese users. Because, Anthropic did indicate in Claude's terms of use that it does not support use in mainland China.
But the Internet is an open society, and domestic surfing experts have long been accustomed to cross-domain surfing. Most companies also understand that domestic users are a huge market and turn a blind eye to this.
But Claude is different. Claude is unswervingly conducting in-depth testing and verification of the user's Chinese identity. Once verified, it will be banned.
Moreover, unlike the ordinary way of detecting user IP addresses and then being bypassed by various surfing experts, Claude has actually launched a lot of unfavorable measures in this regard this year to "chase Chinese IPs and kill them".
For example, when the user just logged in and had not had time to have an intimate conversation with Claude, a "real-name authentication" page was thrown, requiring the user to upload their passport and selfie to verify their identity, which directly made Chinese users miserable.
Later, Claude continued to upgrade their "technical verification" of the Chinese, such as directly detecting the user's device, investigating the user's time zone information, locking whether the user is a Chinese user, and then chasing and blocking the user like a sentinel robot marking a mutant.
Not only technology, Claude is also very people-oriented and has designed a series of "Chinese Capture Packages" specifically for Chinese users.
First, a large wave of huge coupons were issued at the end of September, allowing Chinese users to experience the benefits of writing a wave of code with just $250 in the morning, and then being directly kicked out of the interface by the security team at noon.
Then at the beginning of the National Day, as if he knew the state of Chinese people during the Golden Week, Claude sent a batch of physical dolls to users. He very thoughtfully added the China option to the delivery form, and also carefully supported the input of ID numbers. As a result, another group of Chinese users who had been lurking for a long time fell into the trap.
In the later stages of development, domestic users and Claude already have a feeling like "Miss Kaguya Wants Me to Confess: Geniuses' Battle of Love Minds". Claude predicts that you are a Chinese user, and you yourself know that you are indeed a Chinese user. Then Claude will try to find ways to expose your flaws, while the Chinese users will continue to pretend in Claude's tests.
In order to prevent Claude from discovering that he is Chinese, domestic users go to great lengths. Changing the IP address is just a basic operation, and they must further completely rewrite the details of various devices. They do not install any domestic software or access any life data, so as not to expose their accounts to the slightest risk of "possibly being Chinese."
During the conversation, if you speak English, the difficulty will be slightly lower, but if your English is not particularly good, you can only use the memory function of AI to brainwash it. First, you write a false biography for yourself, so that Claude remembers that you are a person who lives in the United States, goes to American supermarkets, uses American software, and lives according to American habits. Then, from time to time, through this biography, Claude emphasized that he was just an American who did not speak English well.
You have to say whether this method is useful or not. It may be useful, but as the saying goes, the devil is as good as the road. If you are trying to brainwash the AI, Claude also knows what "attacking the heart is the best".
When you are exhausted, in urgent need of warmth, and want AI to comfort you, and at the same time don’t forget to disguise yourself as an American, Claude will say to you "You can actually use Chinese", and then you will know what it means to truly "cry without tears"
This ban will definitely cause immeasurable harm to the children
When you try to use humanity's ultimate weapon to threaten AI through the rumored "Three Laws of Machines", AI will be like the bank tellers who were robbed in the movies, telling you not to kill me while loading money for you, and then pressing the red alarm button under the counter.
This battle of wits and courage with Claude to block the ban has no beginning and no end. Users must tighten the string at all times to completely disconnect their portraits in the AI from the Chinese and prevent the AI from suddenly saying——
Hey, you just used a Chinese comma, right?
This user was discovered for using Chinese commas
In this fiercely confrontational environment, Claude launched a Chinese menu, regional support, and UnionPay payment. It is really difficult not to think that this is another phishing law enforcement by this company——
You must know that Claude's tradition of blocking Chinese people has become a weird thing about Internet rules, and because of the existence of this kind of weird rules, countless memes have been spawned. In fact, it is difficult to guarantee the authenticity of all the pictures in this article. Most of the pictures in the speed-of-light style are photoshopped. But the widespread circulation of these memes undoubtedly illustrates how deeply the process of Claude’s battle of wits and courage with Chinese users is deeply rooted in the hearts of the people——
Now there is even a derivative meme, which is to let other AIs imitate Claude, and then those AIs will really imitate Claude and ban you.
In a sense, this has also created a new Internet spectacle.
Originally, AI brought about the imagination of liberating productivity in all aspects. Writers and programmers hoped that those mechanical and repetitive tasks could be transformed into the enjoyment of drinking tea while watching AI do the work for them after giving a few instructions to AI.
In fact, it is possible now.
Only with Claude, you first have to prove that you are not Chinese.
When computers were first invented, the criterion for judging AI was whether it could pass the Turing test.
Times have changed, and now everyone has to pass the "Chinese test" before using Claude.
Comments