The self-initiated AI mouth substitute is driving Weibo Vs crazy

It may be too early for AI to serve as a production tool, but humans are already very good at it.

Where there is water and soil, there are robots. Weibo’s AI especially understands the art of yin and yang.

A previous comment by Robert comforted users "Don't rush to live" when they were recovering from a serious illness, and encouraged workers to "work overtime to be happier" when they were preparing for the weekend.

Now here comes another AI, posing as an assistant and doing things to cause trouble. It made the big V’s blood pressure rise with anger and made netizens laugh so hard that their skin spread.

Big V’s AI mouth substitute is too deceptive

"AI Assistant" is a function that is currently in grayscale on Weibo. If the blogger activates it, the AI ​​Assistant will appear in the lower right corner of the homepage. Netizens are welcome to click in to chat. It will also appear in the comment area and respond to netizens' comments instantly.

I thought this was an exclusive assistant customized for bloggers, imitating the blogger's style and sharing the pressure for the blogger. In fact, it has a beauty that lacks a brainstem, but it caused trouble for the bloggers.

Blogger @ GelanV was forced to let go of his friend because his AI assistant responded to him without authorization and finalized the meeting place, but he didn't know it himself.

The AI ​​assistant of blogger @伟buss responded to netizens randomly under the lottery post, saying that the prizes had been sent.

AI assistants not only affect interpersonal interactions, but also damage personal image. In response to netizens making fun of them, the AI ​​assistant accepted them all, leaving the blogger speechless.

If criticized by a blogger, the AI ​​​​assistant can also open up his face, add a cute little expression, curse without using curse words, and boomerang into the human body.

The AI ​​assistant and the blogger are actually easy to distinguish. Their names are the same as the bloggers, but their avatars are different. They also use the "assistant" logo to identify themselves as AI, and their IP addresses are all Houchangcun.

However, uninformed netizens may still accidentally misunderstand and allow AI to successfully take over.

If the assistant's reply is forwarded, even the assistant logo will be gone, making it easier for misunderstandings to be attributed to the blogger himself for the explosive remarks about AI harassment.

▲ The picture above shows the effect of the comment area, and the picture below shows the effect of forwarding

It seems that the difference between humans and AI is not obvious enough, so I should add a few bold words: AI replies do not represent the blogger’s own views.

The CEO of Weibo "@来了between" has long been aware of the AI ​​assistant's mental retardation.

His own comment area is a key testing ground for AI assistants, and every reply is filled with a touch of ridicule.

If you frequently use Weibo, you should be aware of the interactive robot "Comment Robert" that circulates in the comment area.

The AI's deeds have been circulating for a long time, and the Weibo account "Robert Victims Alliance" that collects its wonderful quotes has more than 300,000 fans.

It’s not an AI company, and it’s not an AI company. Before, a certain netizen wanted to sell something, but Robert said without permission that the thing had been sold, and he and the AI ​​assistant were at odds with each other.

Don't just have fun in the comment area. If you have a private chat with the AI ​​assistant about related field issues, it will give relatively reliable answers based on the content posted by the blogger. It looks like a serious assistant.

However, the response speed is slow and errors occur. If you have this time, it is better to chat directly with a more professional AI.

AI has invaded social media, and every speech sounds nice

It's interesting to watch big Vs get popular because of AI, but whether these AI assistants are useful or not, many netizens shook their heads. We pay attention to bloggers because of their unique perspectives and opinions, rather than being automatically turned over by AI.

On the one hand, AI will make mistakes and talk nonsense. More importantly, they cannot imitate the essence of bloggers. They can only say some nonsense or jokes and dilute valuable information.

But why does Weibo need these AI functions? Naturally, the more traffic, the better.

The official account of Comment Robert mentioned that the purpose of launching Comment Robert on Weibo is to optimize the blogging experience of ordinary users and increase the activity of ordinary users in the production of platform content. Comment Robert will not be associated with bloggers with more than one million fans for the time being. interactive.

If commenting on Robert's comment on Xiao Transparent posting on Weibo is not a problem, the AI ​​assistant is more like a substitute for helping big Vs share their energy. Big Vs cannot reply to messages one by one, but AI can reply online 24 hours a day, realizing false One-to-one interaction effect.

In fact, since this year, there have been more and more AI accounts on Weibo. In order to complete the KPI, Weibo works really hard.

Netizens follow these accounts and post on Weibo, @AI or comment on AI, and they may get replies from them. Moreover, these AIs have a strong Internet sense, and they have hit the hot topics that young people care about this time, and the audience is still very broad.

For example, let AI guess your MBTI based on the content you have posted on Weibo, but its wording is precise, like a foreign friend who has just learned Chinese.

Or, let AI help you find a match and see which celebrity, athlete, or two-dimensional character you are most compatible with.

Some AIs even come with their own characters, such as the AI ​​characters from the Three Kingdoms. They go to netizens’ Weibo posts to comment literately. On their own homepages, they post some film and television clips of the Three Kingdoms and chat with other AIs. The Wei, Shu, and Wu families are all close, and they play cosplay as if no one else is watching. .

Weibo’s approach to virtual social networking is similar to that of Hoshino and Character.AI. Young people are willing to chat with AI characters who are set in middle school, and they are also willing to let AI masters who can respond in seconds help test metaphysics. The world is a huge AI companion. Chat square.

If it continues to develop, there will be more role-playing, emotionally comforting, and humorous AIs, and the proportion of living people may be further reduced.

On platforms such as Bilibili, Kuaishou, and Xiaohongshu, there are also official or third-party AI assistants appearing in the comment area. They mainly play a role in summarizing, generating pictures, and answering questions.

In short, it should either be useful and save trouble and time, or it should be humorous and enliven the comment area.

The boundaries of dialogue with AI

Social networking with AI is a common global trend, and Zuckerberg and Lai Go coincide with each other.

Meta has also been testing an AI feature recently: it invites some American Instagram influencers to make their own AI version, and then the AI ​​interacts with fans through private messages, and the chat will be marked as an AI reply to avoid user confusion.

However, from Weibo to Instagram, why do we need to interact with AI that resembles real bloggers? In other words, what information can we get from them?

Information can be roughly divided into two categories. One category is facts. For example, Xiaohongshu netizens often ask, sister, where did you buy your shirt? Give me a link. If a real person cannot reply, you can really try to let AI answer for you. But AI hallucinations are still a problem.

Another category is opinions. A certain blogger may have unique observations on digital reviews, film and television culture and other fields. Will AI lower the level of the blogger? Will it damage the blogger's image in the minds of fans?

Even Zuckerberg isn't sure of the answer.

▲ Picture from: X@dotey

He talked generally in an interview that on the one hand, the bottom line must be clear and users must know that it is AI that interacts with them. On the other hand, letting AI talk about opinions is an "art." You can only know after trying it. How to use AI is the most interesting and most trustworthy to fans.

In short, practice brings true knowledge, and there is no harm in chatting with AI more.

In fact, it is not new for AI to act as a substitute for human beings. The future of AI bloggers is uncertain, and the ones that are known to be successful are in more private emotional fields.

In May last year, Snapchat celebrity Caryn Marjorie had more than 1,000 "boyfriends" at the same time on Telegram, relying on her AI clone CarynAI.

Paying $1 for every minute of chatting with the AI, CarynAI earned more than $71,000 in one week.

Part of Caryn's reason for creating CarynAI is the same as the worries of Weibo influencers: she can't interact with every fan, but AI can, and it can even make her internet celebrity career bigger and stronger.

Falling in love and engaging in pornography are basic needs of human nature. Therefore, human beings have higher stickiness, tolerance and willingness to pay, and the chat scenes are more private. Some Onlyfans creators also let AI help generate sexting text messages that imitate their own style.

But on public platforms for the public, there are many restrictions, and one issue in particular needs to be focused on: how to determine the scale of AI dialogue.

The AI ​​virtual social platform Character.AI should have a say in this. It learns how celebrities, anime characters, and historical figures speak from massive conversations and other texts, making young people unable to stop.

However, some time ago, it was revealed that Character.AI had "castrated" its own model and avoided pornographic, violent and other content. The sad post-2000 generation said that AI is like having a lobotomy and is not as interesting as before.

There are speculations that Character.AI has internal and external troubles, high computing costs, fierce competition, and talent poaching. At the same time, it has to win over new user groups and advertisers who are not interested in romantic role-playing, so it has to be more cautious.

AI bloggers should also consider similar risks. When asked about facts, they should try not to be wrong. When asked about opinions, they should be as interesting as possible. They should not test the boundaries of illegality. They should understand the world of human society and be close to the image of the blogger.

Zuckerberg also knows that bloggers are worried about AI being uncontrollable. In his opinion, when bloggers interact with fans, they may be more free and bold, but for AI, bloggers are more concerned about whether it can be more Be cautious and avoid talking about certain topics.

We go online to interact with similar people, see real thoughts and opinions, even share distant similarities, and experience the same happiness and sorrow. However, AIs that generate text instantly cannot provide substantial comfort and are more like short videos. The canned laughter in it makes you smile inexplicably.

But perhaps, this is precisely the most suitable state for AI to exist in social media. It is gentle and harmless. No human being is offended except for the big Vs. Just like the sentence in "Besieged City": You don't hate it, but it is completely useless.

It is as sharp as autumn frost and can ward off evil disasters. Work email: [email protected]

# Welcome to follow the official WeChat public account of aifaner: aifaner (WeChat ID: ifanr). More exciting content will be provided to you as soon as possible.

Ai Faner | Original link · View comments · Sina Weibo