Users use hidden poems to recommend stocks, and Xiaohongshu blocks financial black and gray industries

Economic Observer Follow 2026-09-16 16:30

Economic Observer reporter Chen Yueqin

On September 15th, Xiaohongshu brought the team of community, commercialization, and security governance to the forefront for the first time, showcasing the phased achievements of the platform in regulating financial marketing behavior and addressing financial black and grey industries in the past two years to regulatory authorities such as the China Securities Regulatory Commission, Shanghai Stock Exchange, and multiple licensed securities firms.

With a monthly active user base exceeding 400 million, Xiaohongshu is no longer just a lifestyle community. According to He An, the person in charge of business and financial content operation in Xiaohongshu Community, in the past two years, the daily active users (DAU) of Xiaohongshu Finance have increased by 65%, and the number of financial related searches has increased by 314%.

It is reported that users of Xiaohongshu who are concerned about financial management and finance have three characteristics: firstly, they are mainly from high tier cities, with over 60% of users from first and second tier cities; Secondly, there is a trend towards youthfulness, with over 70% of the population under the age of 40; Thirdly, these users often search with specific questions. 75% of monthly active financial users will search for financial related issues on Xiaohongshu, and 53.5% of financial search terms are long tail keywords with more than 8 words, such as "Which is good to buy medical insurance for millions", "Amendment to the Housing Provident Fund Management Regulations", "Golden breakthrough of $4400", "How to pay flexible employment social security", etc.

The increase in user activity in the financial sector has also attracted the attention of regulatory authorities for illegal traffic diversion and stock recommendations on social media platforms. In April of this year, eight departments including the People's Bank of China and the Ministry of Industry and Information Technology jointly issued the "Management Measures for Online Marketing of Financial Products", which will come into effect on September 30, 2026.

Xiaohongshu conducted a self-examination in accordance with regulatory requirements and updated its community rules. If the rules are too strict, creators and institutions will complain that normal science popularization cannot be released; but if there are omissions, there will be illegal diversion and black and gray production problems. "Huilin, the compliance officer of Xiaohongshu Finance, pointed out that a good ecology requires joint construction by all parties. This is the background of Xiaohongshu's decision to make the rules transparent and release the first" Community Financial Ecology Convention ".

The concentrated risks faced by Xiaohongshu in the field of financial governance include illegal stock recommendations, illegal overseas securities exhibitions, illegal lending, as well as high leverage trading such as foreign exchange and virtual currencies.

The governance team of Xiaohongshu also continues to compete with users who violate regulations such as illegal stock recommendations. In the early days, violators directly posted profit charts in their notes, accompanied by keywords such as "eating meat", "continuous board trading", and "buying followed by continuous gains", guiding users to privately chat and redirect traffic outside the platform. After being discovered and severely punished by the platform, these users constantly upgrade their language to evade the platform's review mechanism, such as disguising stock recommendation information as "express delivery pick-up code", "library seat search", "subway station guide" and other daily scenarios, hiding the recommended stock codes for the day in a string of numerical combinations. Some users use hidden poems to recommend stocks or disguise themselves as ordinary users, posting "stock trading daily" to deceive system traffic, and then redirect users outside the website.

In addition to the disguise in language, the black and gray industry has also formed a highly organized "one-stop" service in the falsification of financial marketing qualifications. In the certification process of professional accounts, Black and Grey Industry obtains real license screenshots through internal employees of financial institutions, or resells scarce license resources through registering shell companies and transferring equity.

The governance manager of Xiaohongshu Financial Professional Account, Kaka, mentioned that its core KPIs (key performance indicators) mainly consist of two points: not allowing people who should not have obtained their identities to obtain them, and not allowing those who have obtained their identities to do things they should not have done.

Kaka revealed that there are also vulnerabilities in facial recognition technology that can be exploited by black and gray industries. Part of the black industry is using AI face swapping technology, such as replacing the opening and blinking of legal entities, in an attempt to break through the live detection process of the system. When submitting the business license review, they will also use the same template to tamper with the license text, seal position, or QR code.

The governance of Xiaohongshu mainly consists of three steps: risk identification, content recognition, and graded disposal. In terms of identifying risks, in addition to manual and public opinion perception, the governance team will also disguise themselves as ordinary users to test and see if criminals will actively engage. However, with the continuous upgrading of black and gray production technology, traditional keyword interception and manual review are no longer able to fully identify risks.

We have introduced AI big models for comprehensive judgment when dealing with users' covert violations. ”Hansen, the head of financial content governance at Xiaohongshu Community, introduced that based on the cross scenario understanding ability of the big model, the system can combine multiple dimensions such as notes, comments, and private messages to restore users' true intentions.

For example, when a black industry uses hidden poems to recommend stocks, traditional systems are difficult to identify, but by combining the user's traffic behavior in the comment section with a large model, violations can be accurately determined; For black and gray production gangs with hundreds or thousands of accounts cooperating with each other and using small accounts as "childcare" in the comment section, the governance team will identify and crack down on them in batches through WiFi, IP, group chat, account relationships, and behavioral similarity.

In the past six months, Xiaohongshu has disposed of 157000 illegal accounts, over 1.01 million illegal notes, and over 4.77 million comments.

The Community Financial Ecology Convention is divided into three parts: financial content management standards, financial marketing content management standards, and violation disposal. Among them, in terms of content standards, Xiaohongshu has imposed strict restrictions on misleading information, such as absolute language such as low threshold, instant payment, and guaranteed capital and interest; In terms of marketing subject qualifications, institutions without financial licenses are strictly prohibited from entering; Financial practitioners must upload valid qualifications and proof of employment within the past month in order to carry out financial marketing. Users who engage in illegal financial activities such as recommending stocks will be banned for the first time, and their accounts will be suspended if they do so again.

Eight types of behaviors, including malicious criticism, illegal stock recommendations, illegal financial activities, forging authoritative endorsements, promising to "guarantee principal and interest", fabricating false investment profiles, and disguised diversion outside the station, are explicitly prohibited.


Can be contacted: chenyueqin@eeo.