The data show that 800 million comments are produced every day, and 32.79% of them are "tiktok and personal attack" in the improper comment. Recently, tiktok has further upgraded the ability of platform to prevent network bursts. Among them, for inappropriate comments, tiktok will be sent on-line to warn the function and intercept. Tiktok will be on line with "risk warning" and "negative content filtering" for inappropriate private credit. Previously, before the Spring Festival, tiktok targeted special search teams such as human flesh search, privacy, humiliation and abuse. Special network governance teams were set up to focus on social networking and other Internet hotspots to stimulate Internet users to engage in violent human activities, such as human flesh search, abuse and attack. At that time, the trembling tiktok launched two measures from the perspective of victim treatment and victim protection. In the aspect of harnessing tiktok, the main reason is to enhance the inspection and control by focusing on the upgrading of governance models, and concentrate on the regulation of network violence through the commentary area or private letter. The data show that 800 million comments are produced every day, and 32.79% of them are "tiktok and personal attack" in the improper comment. Massive comments, if only handled after the event, will undoubtedly make the platform face great governance pressure. More importantly, online violence has occurred. And the tiktok will launch the "warning" function, which is to pre - defense the ability. Specifically, based on this function, if the system model determines that there is a problem with the comments sent by the user, it will push a pop-up reminder and suggest the sender to edit again to remove improper remarks. If the user who receives the pop-up reminder insists on sending, the comment will not appear in the comment area. It is reported that in the recent small-scale test, the function has reduced about 30000 improper comments per day. In addition, at the commentary side, the tiktok has three functions: "comment hierarchy management", "keyword shielding" and "comment selection". Among them, when using "comment permission hierarchy", users can set comment permission through "permission settings - Advanced Settings - who can comment". According to their own needs, users can choose who can comment on their published content from the four options of "everyone", "only fans", "only mutual friends" and "only themselves". For private content, the two functions of "tiktok" are added to "user risk reminders" and "negative content filtering". Under the "user Risk Reminder" function, when the system model identifies that there is a risk in the private letter of a stranger just received by the user, a pop-up window will be pushed to the user to prompt him to "know the user". If the user selects "don't know", the system will pop up again and prompt the user to "blackmail, report, remove fans" and other operations. Under the "negative content filtering" function, when the model determines that the private message received by the user is "abuse or harassment", the message will be directly put into the "stranger box - spam". The data of shaking tiktok show that the function has been daily service of 70 thousand users, reducing the negative information they receive. (Xinhua News Agency)