Censorship by TikTok
Internal guidelines depending on the country against the promotion of violence, separatism, and "demonization of countries" could be used to prohibit content related to the 1989 Tiananmen Square protests and massacre, Falun Gong, Tibet, Taiwan, Chechnya, Northern Ireland, the Cambodian genocide, the 1998 Indonesian riots, Kurdish nationalism, ethnic conflicts between blacks and whites or between different Islamic sects.
A more specific list banned criticism of world leaders, including past and present ones from Russia, the United States, Japan, North and South Korea, India, Indonesia, and Turkey.
[1][2] In September 2019, The Washington Post reported allegations from former U.S. employees that TikTok censored content sensitive for Beijing as well as political discussions unrelated to China.
[3][4] On 27 November 2019, TikTok temporarily suspended the account of Feroza Aziz after she posted a video (disguised as a makeup tutorial) which drew attention to the Xinjiang internment camps.
[8][9] TikTok's policies ban content related to a specific list of foreign leaders such as Vladimir Putin, Donald Trump, Barack Obama, and Mahatma Gandhi because it can stir controversy and attacks on political views.
[12] In March 2020, internal documents leaked to The Intercept revealed that moderators had been instructed to censor political speech in livestreams, banning those who harmed "national honor" or who broadcast streams about "state organs such as police".
[16][17][18] In June 2020, The Wall Street Journal reported that some previously non-political TikTok users were airing pro-Beijing views for the explicit purpose of boosting subscribers and avoiding shadow bans.
It stated that although this and other loopholes were patched by TikTok before the end of March, the initial failure to correctly implement the restrictions, in addition to the effects from Kremlin's "fake news" laws, contributed to the formation of a "splInternet ... dominated by pro-war content" in Russia.
[35][36] A 2023 paper by the Internet Governance Project at Georgia Institute of Technology concluded that TikTok is "not exporting censorship, either directly by blocking material, or indirectly via its recommendation algorithm.
[38] In May, the Acton Institute was suspended after it promoted videos about the imprisonment of Jimmy Lai and the Chinese government's crackdown on the pro-democracy camp in Hong Kong.
[50] In 2019, The Guardian reported that TikTok's efforts to provide locally-sensitive moderation had resulted in the removal of content that could be perceived as being positive towards LGBTQ+ people or LGBTQ+ rights (such as same-sex couples holding hands) in countries such as Turkey.
[52] TikTok has since apologized and instituted a ban against anti-LGBTQ ideology, with the exceptions of places such as China, the Middle East, and parts of Europe where additional censorship laws may apply.
TikTok released a statement apologizing for this, claiming that a technical glitch had caused the display error and that the hashtags had received over 2 billion views.
[59] Hicks argued that LGBTQ+ people and person of color have found that the guidelines are enforced "wildly differently", meaning their content will be suppressed or removed for supposed violations and that reports of harassment from other users are not acted upon: "Not only is it hurting their ability to speak and be seen on the app, but it's also allowing them to get attacked and have hate speech thrown their way.
"[52] He told CNN that he welcomed TikTok's public pledge of support to the Black community after the 2020 police killing of George Floyd and that he applied to the company because he felt its corporate value "really resonated with me.
[61] In 2021, TikTok apologized and vowed to do better after an app called for black creators to be treated more fairly amid accusations of censorship and content suppression was suspended.
TikTok has community guidelines, but there is no public list of specific words and phrases that are banned, and it is not clear how much moderation is done algorithmically versus by actual people.
