news

HowNet, which has been silent for a long time, has launched a strong attack on AI.

2024-08-20

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina


HowNet is causing trouble again

Just two days ago, Shanghai Mita Network Technology Co., Ltd. issued a statement saying that it had received an infringement notification letter on the 15th.

This notice is a lengthy 28-page letter, and the simplified version is just one sentence:Mita AI search can retrieve the titles and abstracts of my academic literature. Is this infringing?” 。


The purpose is actually very obvious, and it is clearly written in the original text:If you need business cooperation, please contact us” 。

This is not right,Send moneyDo you mean?

And who is this "China Academic Journal (CD-ROM Edition)" Electronic Magazine Co., Ltd.?

Upon closer inspection, it turned out to be our old friend CNKI, which is familiar to everyone.

This man who once single-handedly defeated "Dr. Zhai" was suspended by the Chinese Academy of Sciences in the past two years, sued by Professor Zhao Dexin, and even fined more than 80 million yuan by the State Administration for Market Regulation for anti-monopoly. Now he turns around and reaches out to AI again?

The Secret Tower is not afraid either.Directly respond in the articleThe abstracts and titles of the documents are not your big treasures that can make money. After our AI collected them, it honestly posted the link to your CNKI, and there is no infringement at all.


They even said: "Without search, there is no research" and criticized CNKI: "If scientific literature becomes a luxury (like CNKI), it will be detrimental to the fair acquisition of knowledge and the development of scientific research."


Even on the official website interface, there is a spoof sentence "Damn! We received a 28-page infringement notice from CNKI” 。


But no matter what, Secret Tower may have cut off its relationship with Lightspeed in order to avoid risks.


Because of this wave of events, we even received olive branches from several other database companies.


We immediately tried the Mita search and found that their current paper sources include the Internet,Basically, it is based on the content of Wanfang database.



We first consulted the company's legal department, and he told us,The most important thing is to see whether the acquisition of information involves profit-making nature.


On the other hand, we also found our friend A Tian (pseudonym), who is a doctoral student in law at Tsinghua University, and had a chat with him.

He told us that the whole thing was very strange. In his opinion, CNKI was most likely trying to scare the other party. If it really came to a lawsuit, no matter how the court ruled, he thought it was very likely that CNKI would not win.

A Tian told us that the so-called lawyer's letter,Nowadays it is often used as a "legal strategy", to put it bluntly, is to scare the other party.


For example, when we surf the Internet, we often see celebrities encountering some melodramatic farces, and the first thing they do is send a lawyer's letter.

But the fact that he sent a lawyer's letter does not mean that this bloody farce has been whitewashed by the law. Many times, he just wanted to scare people.

So, will CNKI's lawyer's letter to Mita be legally viable?It is not CNKI that has the final say, it is the court that has the final say

Moreover, A Tian feels that the legal basis for CNKI in this matter is not sufficient.

Because the data used by Secret Tower is actually the title and abstract, which can be seen on the CNKI page by just searching, even without logging in.To put it bluntly, it is actually public.Therefore, there is nothing wrong with the Secret Tower AI searching and retrieving this part of the content.

The abstract and title of the paper are completely public in CNKI.


Moreover, in A Tian’s opinion, although many domestic papers, especially core journals, are included in CNKI, they are actually also made public for free on the journal’s official website, official WeChat account and other platforms. If AI captures papers through these channels, even though they are the same as those included in CNKI, it is difficult to say that AI has infringed CNKI. . .


Unless the article included in CNKI itself is a paid non-public resource, and then AI search uses technology to crack the full text and make the content public, that would be infringement.

But there may also be some improper operations in the secret tower.

According to Jiemian News,Perhaps the podcast and library sections searched by Mita AI may have index libraries.

Image fromAIProductRena


That is to say, the Secret Tower first built a "reservoir" inside for the documents it collected in batches.

When users search, MiTa will search for fresh information externally, and then integrate this fresh data with the content in the "reservoir" to provide answers.

If this operation is carried out, you may face legal risks.

We also tried to let the AI ​​of Mita answer the question by itself, and it turned out that it was indeed a little unconfident.


Is it true that Zhiwang is causing trouble for MiTa purely because of copyright?

I can only say that it is not certain.

Because whether the search engine can crawl the website content is determined by the website robots file settings.

If the robots file of the website is not allowed to be crawled, but the search engine forcibly crawls it, then infringement is involved.

However, the robots file page of CNKI main site does not prohibit any search engine crawlers., then how can we say that the secret tower violated the rules?


So, comprehensively speaking,The reason why Secret Tower temporarily disconnected from CNKI is probably for safety reasons, while CNKI is more likely to protect its own interests.

You may not be very familiar with MiTa. MiTa Technology was founded in April 2018 and took off immediately after launching AI search.

In March this year, according to SimilarWeb data, the average daily visits to the "new generation" MiTa AI search website reached more than 200,000 times, with a monthly growth rate of 551.35%.

Our editorial department discovered the Secret Tower AI Search at the beginning of this year. At that time, it relied on its excellent information retrieval capabilities, especially the "in-depth" search mode, to help us quickly find the information we want from the vast number of papers, and it quickly became popular in the editorial department.


Later, the Secret Tower became popular, and even not long ago,Mita has just completed a new round of financing of over 100 million yuan, the valuation has risen to $150 million.

Now that you have money in your pocket, CNKI comes to give you a reminder that it seems to be natural to want to cooperate with MiTa to make some money.

But unexpectedly, Secret Tower was extremely tough and did not give CNKI any face at all.

Although it has gained everyone's attention and sympathy, the road ahead for Mita may be very difficult. Because even though Mita has just received financing, the competition among these new AI forces has reached an unimaginable level.

I would like to ask those friends who often visit B station,Have you been brainwashed by Kimi's overwhelming advertisements in the first half of the year?

Image from 36Kr


With this massive marketing, Kimi has achieved quite obvious results, with a tendency to break out of the circle.

Image from Intelligent Emergence


Among this group of new AI forces, MiTa's biggest feature is its excellent deep search. Now that it has cut off its connection with CNKI, it remains to be seen whether it can maintain its own characteristics.

We are also a little surprised about CNKI's operation this time.

It is clear that AI searches such as MiTa are, to a certain extent, used to divert traffic to them. For example, when I use them myself, I often click directly into the CNKI official website from the reference source, and then log in to read the full text.

But they chose to use this method that is obviously not accepted by the outside world to play a big game...

But no matter what, the story of CNKI and Secret Tower is also a reminder for everyone in advance.

In the future, the confusion in AI data applications will probably bring more controversy.

Just last month, Condé Nast, which owns media outlets such as The New Yorker, Vogue and Wired, sent a similar cease-and-desist letter to the overseas secret agency Perplexity, also accusing the AI ​​search company of plagiarism.

A month earlier, Forbes also made the same accusation against Perplexity.

This time, CNKI may have made a big deal out of nothing, but what about next time?

Written by: Bajie

edit: Jiang Jiang & Noodles

Art: Xuanxuan

Image, source

Engader:Condé Nast has reportedly accused AI search startup Perplexity of plagiarism

Axios: Scoop: Forbes threatens Perplexity with legal action

Forbs:Why Perplexity’s Cynical Theft Represents Everything That Could Go Wrong With AI

Jiemian News: The dispute between CNKI and Metadata: Where is the copyright boundary of AI search engines?

Secret Tower

CNKI official website

Jiazi Guangnian: After disassembling SearchGPT, we discovered the barriers, breakthroughs, and future of AI search