
Douyin plans to make its algorithm public

Respond to the controversy with concrete actions
Author | Liu Baodan
Editor | Zhou Zhiyu
More than a month ago, Zhong Shanshan called out Zhang Yiming, criticizing Douyin's algorithm for playing with the intelligence of the public. A month later, Douyin responded with concrete actions.
On January 3rd, Wall Street News learned that Douyin announced the launch of 10 measures to establish a safety and trust center, effectively promoting the transparency of algorithms and platform governance, creating a safe and trustworthy platform environment, and building a better online ecosystem.
This means that Douyin, which has 700 million daily active users, is beginning to actively unveil the mystery of internet algorithms.
Earlier, Meituan, Xiaohongshu, and others also took related measures. On December 27th, Meituan announced eight algorithm improvement measures, including normalizing algorithm transparency and diversifying algorithm communication mechanisms. On November 21st, Xiaohongshu Vice President Xu Lei stated that based on personalized recommendation algorithms, Xiaohongshu would allocate more than half of its traffic exposure to users with fewer than 1,000 followers, practicing traffic inclusivity.
At the beginning of last month, Douyin Vice President Li Liang responded to algorithm issues, stating, "All major platforms use algorithms; they are not as complex and mysterious as they seem. However, the popularization of algorithms is inadequate, transparency is insufficient, and the identification and crackdown on rumors and online violence are not strong enough. This is indeed something that our platform needs to strengthen."
Now, Douyin has officially put algorithm transparency on the agenda, focusing primarily on issues such as rumor governance and information cocoons.
Douyin stated that it will rely on the safety and trust center to focus on socially concerned topics, including algorithm principles, review mechanisms, combating online violence and rumors, platform governance work, and protection for minors and the elderly. "We hope to make the public feel more connected to the platform's work, respond to public concerns, and accept public supervision."
Regarding the controversial issue of online rumors, Douyin stated that it will comprehensively upgrade its service capabilities, especially for parties involved in hot topics and online violence, providing "one-stop services" and pursuing infringement issues "to the end," addressing pain points such as "slow response to infringement issues" and "inadequate governance."
Douyin will also launch special services for "enterprise-related infringements," efficiently resolving reasonable demands, strictly controlling rumor and online violence, proactively managing infringement information, and legally resolving infringement issues to the maximum extent.
Douyin stated that it will strengthen product capability construction and enrich the existing "one-click anti-online violence" intelligent blocking function. In addition, Douyin will expand cooperation with authoritative legal aid organizations to provide legal consultation assistance and psychological counseling services to users in need.
Douyin claimed that it will enhance the application of algorithms in rumor governance, accurately recommending debunking content to users who have browsed false information. At the same time, for controversial content, it will expand cooperation with authoritative media and organizations in the "debunking team" and strengthen manual assessment and verification in various hot events. Douyin will also launch a video "tagging" function, encouraging discerning users to participate in "tagging" controversial content to improve users' ability to discern information.
In fact, in May 2024, Douyin began implementing a verification mechanism for hot content. For users who speak out as parties involved in hot events on Douyin, the platform will contact them for identity verification For the interpretive content published by creators, Douyin requires publishers to prominently label it with a "fictional interpretation" statement. For some suspected staged trending content, Douyin stated that it will request relevant local departments to verify or collaborate with local news agencies for investigation.
In response to societal concerns about the "information cocoon" issue, Douyin has also provided specific governance ideas.
Douyin stated that it will further improve its recommendation algorithm by enhancing mechanisms such as content diversity exploration, interest exploration and matching, search recommendation linkage, and social friend recommendations, supplementing various recommendation signal sources to meet users' long-tail interests as much as possible. It will focus on solving the problem of "content consumption density" by increasing duplicate content filtering.
In addition, Douyin will further improve the settings for recommendation blocking words and the "not interested" function, giving users better management capabilities for recommended content.
At the same time, Douyin will increase the promotion of high-quality content. In the past two years, Douyin has added a separate "high-quality" sub-link in its existing recommendation chain and included signals such as "user favorites and revisits, long-term companionship from authors, and a high-quality community atmosphere" as indicators for discovering high-quality content.
According to the plan, by 2025, Douyin will further enhance its support mechanisms for high-quality content, expecting to increase the exposure of high-quality content by 300%.
Douyin is attempting to establish a more open communication mechanism and has provided specific measures.
Douyin stated that in the future, it will actively listen to suggestions from all sectors through user research, public and media open days, and accept public supervision. On the Douyin Safety and Trust Center website, a public opinion solicitation section will be established to specifically collect societal opinions on algorithms and platform governance. Additionally, an algorithm consultation hotline will be added to the service hotline 95152 to specifically handle various user supervision and criticism.
Previously, Douyin has already accumulated experience in platform governance. In the past year, Douyin has taken multiple measures to strengthen platform governance, such as addressing the upstream pollution of algorithms caused by accounts teaching illegal "operational techniques" and "account creation guides." In June 2024, Douyin organized a special action to combat content that teaches "bottomless traffic tricks."
Over the past decade, ByteDance's main challenge has been how to grow under the siege of giants like Alibaba and Tencent. In the past two years, ByteDance has firmly established itself and has become a new leader in the internet industry. As its performance steadily grows, ByteDance has also begun to take on more social responsibilities. Publicly disclosing algorithm transparency is just the beginning, and as ByteDance's global influence continues to expand, the responsibilities it needs to bear are also increasing
