This report is also available in Arabic , Chinese , and French .
本报告还提供阿拉伯文、中文和法文版本。
EXECUTIVE SUMMARY 摘要
The People’s Republic of China (PRC) actively attempts to manipulate and dominate global discourse on Xinjiang and to discredit independent sources reporting ongoing genocide and crimes against humanity conducted against predominantly Muslim Uyghurs and members of other ethnic and religious minority groups in the Xinjiang Uyghur Autonomous Region. PRC-directed and -affiliated actors lead a coordinated effort to amplify Beijing’s preferred narratives on Xinjiang, to drown out and marginalize narratives that are critical of the PRC’s repression of Uyghurs, and to harass those critical of the PRC.
中华人民共和国（中国）积极试图操纵和主导全球关于新疆的话语，并诋毁独立消息来源，这些消息来源报道了对新疆维吾尔自治区以穆斯林为主的维吾尔族和其他少数民族和宗教少数群体成员的持续种族灭绝和危害人类罪。 由中国指导和附属的行为者领导了一项协调一致的努力，以扩大北京对新疆的偏好叙事，淹没和边缘化批评中国镇压维吾尔人的叙事，并骚扰那些批评中国的人。
MESSAGING TACTICS 消息传递策略
PRC messaging tactics seek to drown out critical narratives by both flooding the international information environment to limit access to content that contradicts Beijing’s official line, and by creating an artificial appearance of support for PRC policies. Messengers use sophisticated A.I. -generated images to create the appearance of authenticity of fake user profiles. The PRC works to silence dissent by engaging in digital transnational repression, trolling, and cyberbullying.
中国的信息传递策略试图通过充斥国际信息环境以限制对与北京官方路线相矛盾的内容的访问，以及通过制造一种支持中国政策的人为表象来淹没批判性叙事。信使使用复杂的人工智能生成的图像来创建虚假用户配置文件的真实性外观。中国致力于通过数字跨国镇压、拖钓和网络欺凌来压制异议。
Flooding To Drown Out Critical Narratives
洪水淹没了批判性叙事
The PRC floods conversations to drown out messages it perceives as unfavorable to its interests on search engines and social media feeds, and to amplify Beijing’s preferred narratives on its treatment of Uyghurs. Pro-PRC stakeholders flood information ecosystems with counternarratives, conspiracy theories, and unrelated news items to suppress narratives detailing PRC authorities’ atrocities in Xinjiang. Government social media accounts, PRC-affiliated media, private accounts, and bot clusters, likely all directed by PRC authorities, assist in this effort.
中国充斥着大量对话，以淹没它认为不利于其在搜索引擎和社交媒体上的利益的信息，并放大北京在对待维吾尔人方面的首选叙述。亲中利益攸关方用反叙事、阴谋论和不相关的新闻充斥着信息生态系统，以压制详细描述中国当局在新疆暴行的叙述。政府社交媒体账户、中国附属媒体、私人账户和机器人集群，可能都是由中国当局指导的，协助这项工作。
Astroturfing To Create a False Appearance of Support
Astroturfing 以创造虚假的支持外观
To manipulate narratives on Xinjiang, pro-PRC actors engage in “astroturfing ,” or coordinated campaigns of inauthentic posts to create the illusion of widespread grassroots support for a policy, individual, or viewpoint, when no such widespread support exists. Similar to flooding, the PRC uses astroturfing to inundate the information space with “positive stories ” about Xinjiang and the Uyghur population, including manufactured depictions of Uyghurs living “simple happy lives,” as well as posts emphasizing the purported economic gains that the PRC’s policies have brought to Xinjiang. In mid-2021, more than 300 pro-PRC inauthentic accounts posted thousands of videos of Uyghurs seeming to deny abuse in the region and claiming they were “very free.” These videos claimed to show widespread disagreement throughout Xinjiang with claims in international media that Uyghurs were oppressed. However, according to the New York Times and ProPublica , propaganda officials in Xinjiang created most of these videos, which first appeared on PRC-based platforms and then spread to YouTube and Twitter, in order to manipulate public opinion.
为了操纵有关新疆的叙事，亲中行动者进行“造星”，或协调不真实帖子的运动，以制造一种错觉，即基层广泛支持一项政策、个人或观点，而这种广泛支持并不存在。与洪水类似，中国利用占星术，用关于新疆和维吾尔族人口的“正面故事”淹没信息空间，包括对维吾尔人过着“简单幸福生活”的捏造描述，以及强调中国政策给新疆带来的所谓经济收益的帖子。2021年年中，300多个亲中国的虚假账户发布了数千个维吾尔人的视频，这些视频似乎否认在该地区受到虐待，并声称他们“非常自由”。这些视频声称，新疆各地普遍存在分歧，国际媒体声称维吾尔人受到压迫。然而，据《纽约时报》和ProPublica报道，新疆的宣传官员制作了大部分这些视频，这些视频首先出现在中国的平台上，然后传播到YouTube和Twitter，以操纵公众舆论。
A.I. Generated Images Used To Create the Appearance of Authenticity
人工智能 用于创建真实外观的生成图像
Since at least January 2021 , pro-PRC networks have used advanced artificial intelligence-generated content, such as StyleGAN machine-learning generated images, to fabricate realistic-looking profile pictures for their inauthentic accounts. Unlike stolen images of real people, these tools create composite images that cannot be traced using a reverse image search, making it harder to determine whether the account is inauthentic. Some of these accounts repeatedly denied the PRC’s atrocities in Xinjiang, falsely asserting that the body of overwhelming and objective independent evidence of the atrocities is simply a fabrication of the United States and its allies.
至少从 2021 年 1 月开始，亲中国网络就使用先进的人工智能生成内容，例如 StyleGAN 机器学习生成的图像，为其不真实的帐户制作逼真的个人资料图片。与被盗的真人图像不同，这些工具创建的合成图像无法使用反向图像搜索进行追踪，因此更难确定该帐户是否不真实。其中一些说法一再否认中国在新疆的暴行，错误地声称，关于暴行的压倒性和客观的独立证据只是美国及其盟国的捏造。
Transnational Repression, Trolling, and Cyberbullying To Silence Dissent
PRC-sponsored transnational repression targets those who speak out against the PRC, particularly in Chinese diaspora communities , with on- and offline harassment to prevent them from sharing their stories or to intimidate them into self-censorship. Trolling campaigns are designed to silence those who speak out against the PRC, to poison the information environment with bad-faith arguments, and to silence opposing viewpoints. Trolling campaigns frequently evolve into threats of death, rape, or assault; malicious cyber-attacks; and cyberbullying or harassment through doxxing – publishing an individual’s personal information online without their permission, including their full name, home address, or job. In March 2021, the PRC Ministry of Foreign Affairs (MFA) publicly questioned several individuals’ claims of maltreatment.
Narrative Focus
PRC Xinjiang narratives focus on denying criticism and amplifying “positive stories” in an attempt to counter accusations of genocide and crimes against humanity. The most aggressive PRC messengers often go on the offensive, creating false equivalencies with the actions of other countries to distract from international criticism of PRC behavior.
Rebutting/Denying Criticism from Independent Media Sources
PRC messengers both post and amplify content that denies claims made by independent media outlets and internationally renowned think tanks. In response to third-party accusations that the PRC subjects Uyghurs to forced labor , a wave of PRC diplomatic accounts , PRC- and Chinese Communist Party (CCP)-affiliated media organizations , and suspected bot networks posted stories about the mechanized cotton harvesting process in Xinjiang, suggesting that the Xinjiang cotton industry has no need for forced labor. This messaging avoided responding to reports regarding the PRC authorities’ transfer of an estimated 100,000 Uyghurs out of Xinjiang in “coercive labor placements ” to work in factories elsewhere in the PRC.
Amplifying “Positive Stories” To Counter/”Disprove” Accusations of Genocide and Crimes Against Humanity
PRC actors use hashtags such as #AmazingXinjiang and #Xinjiang to amplify positive stories about Xinjiang and counter independent reporting of allegations of crimes against humanity and genocide by PRC authorities. Stories of a multicultural society living in harmony stand in contrast to the reality of the PRC’s extensive surveillance of Uyghurs, including PRC officials living in Uyghur homes for at least six weeks a year. This messaging aims to divert attention from reports regarding the PRC’s “demographic engineering ” campaign to systematically increase the Han Chinese population in Xinjiang and to “dilute ” Uyghur population concentrations in the region.
“Whataboutism” and False Equivalencies Used To Distract/Deflect Criticism
PRC actors, including voluble diplomats in the MFA’s Information Department use “whataboutism” and false equivalencies to distract from the PRC’s policies in Xinjiang and to portray accusers as hypocritical . Their arguments do not advance the case that the PRC is innocent; rather, they make the point that other countries are equally guilty of abuses. Despite these efforts to distract from the situation in Xinjiang, independent media outlets, academics, and human rights activists have published multiple eyewitness accounts and verifiable data that the PRC has imprisoned an estimated one million people and that credible evidence exists of torture , forced sterilization , and other abuses.
中国行为者，包括外交部新闻司的外交官，使用“什么主义”和虚假的等同来分散人们对中国新疆政策的注意力，并将指控者描绘成虚伪的。他们的论点并不能证明中华人民共和国是无辜的;相反，他们指出其他国家也同样犯有侵权行为。尽管有这些努力来转移人们对新疆局势的注意力，但独立媒体、学者和人权活动人士还是公布了多个目击者的说法和可核实的数据，表明中国已经监禁了大约100万人，并且存在酷刑、强迫绝育和其他虐待行为的可靠证据。
PRC MESSENGERS 中国信使
The PRC’s most aggressive messengers are a subset of PRC diplomatic officials known for their confrontational messaging. Additionally, PRC- and CCP-affiliated media spread Xinjiang-related disinformation on a global scale in at least a dozen languages. To reach and resonate with global audiences, the PRC turns to private media companies and multilingual social media influencers. Trolls take the lead on attacking, stirring controversies, insulting, and harassing netizens to poison the information environment and distract from narratives critical of the PRC.
中国最具侵略性的信使是中国外交官员的一部分，他们以对抗性信息而闻名。此外，与中国和中共有关联的媒体以至少十几种语言在全球范围内传播与新疆有关的虚假信息。为了接触全球受众并引起共鸣，中国求助于私营媒体公司和多语种社交媒体影响者。喷子带头攻击、挑起争议、侮辱和骚扰网民，毒害信息环境，分散对批评中华人民共和国的叙事的注意力。
Subset of PRC Diplomats Lead with Assertive Messaging
部分中国外交官以自信的信息为主导
Most of the PRC’s diplomatic social media messaging is positive and tends to focus on highlighting good relations with other countries and seeks to burnish the PRC’s image. A minority of MFA officials – dubbed “wolf warriors ” by some commentators – use social media platforms to defend the PRC’s national interests, often in confrontational ways . These individuals are most likely to try to deny, “disprove,” and deflect narratives that run counter to PRC official messaging. For example, to distract from the atrocities in Xinjiang, PRC messengers spread a false narrative claiming that the CIA was trying to foment unrest in Xinjiang in order to bring down the PRC. This aggressive style allows the PRC to experiment with different types of messaging to see what plays well at home and abroad. For example, some MFA officials’ accounts repeatedly spread disinformation and conspiracy theories regarding the origin of the virus that causes COVID-19 and about Russia’s unjustified and unprovoked war against Ukraine.
中国的大多数外交社交媒体信息都是积极的，往往侧重于强调与其他国家的良好关系，并试图提升中国的形象。少数外交部官员——被一些评论员称为“战狼”——利用社交媒体平台捍卫中国的国家利益，往往以对抗的方式。这些人最有可能试图否认、“反驳”和转移与中国官方信息背道而驰的叙述。例如，为了转移人们对新疆暴行的注意力，中国信使散布虚假叙述，声称中央情报局试图煽动新疆骚乱，以推翻中华人民共和国。这种咄咄逼人的风格使中国能够尝试不同类型的信息传递，看看哪些在国内外都行得通。例如，一些外交部官员的账户反复散布关于导致 COVID-19 的病毒起源以及俄罗斯对乌克兰的无理和无端战争的虚假信息和阴谋论。
PRC- and CCP-Affiliated Media Spread Xinjiang-Related Disinformation Globally
中共和中共下属媒体在全球散布涉疆虚假信息
PRC- and CCP-affiliated media outlets like China Global Television Network, China Daily, China Radio International, and Xinhua produce content in at least 12 languages and devote significant resources to advertising on social media. In February 2021, facing growing international scrutiny over the PRC’s genocide in Xinjiang, Xinhua released a “fact sheet ” containing numerous false claims, such as stating that the internment camps holding Uyghurs in Xinjiang are “vocational education and training centers”’ that have “fully guaranteed the trainees’ personal freedom and dignity.” However, detainees’ testimonies published by Amnesty International allege that the PRC subjected them to regular interrogation, torture, and other mistreatment. The PRC partners with foreign media to republish both PRC-produced and PRC-backed content to local audiences, giving Beijing’s chosen narratives a level of authority and credibility they would not be able to achieve on their own. For example, Kenya Broadcasting Corporation ran a story by an anonymous author in November 2019 on the PRC’s “poverty alleviation ” policy in Xinjiang, causing observers to question its validity and whether it was PRC propaganda.
中国环球电视网、中国日报、中国国际广播电台和新华社等中国和中共附属媒体机构以至少12种语言制作内容，并投入大量资源在社交媒体上做广告。2021年2月，面对国际社会对中国在新疆的种族灭绝行为日益严格的审查，新华社发布了一份“情况说明书”，其中包含许多虚假说法，例如称新疆关押维吾尔人的拘留营是“职业教育培训中心”，“充分保障了受训者的人身自由和尊严”。然而，国际特赦组织公布的被拘留者证词称，中国对他们进行定期审讯、酷刑和其他虐待。中国与外国媒体合作，向当地观众重新发布中国制作和中国支持的内容，使北京选择的叙述具有他们自己无法实现的权威和可信度。例如，肯尼亚广播公司在2019年11月刊登了一位匿名作者关于中国在新疆的“扶贫”政策的报道，引起观察人士质疑其有效性以及是否是中国的宣传。
PRC Increasingly Turns to Private Media Companies To Craft Foreign-Facing Information Manipulation Campaigns
中国越来越多地求助于私营媒体公司来策划面向外国的信息操纵活动
The PRC outsources and privatizes some of its foreign language information operations to take advantage of private sector innovation. The PRC government engages with at least 90 PRC-based firms to design foreign-facing information manipulation campaigns to portray the PRC positively. For example, a publishing organization operated by the Xinjiang Uyghur Autonomous Region Bureau of Radio, Film and Television and affiliated with the CCP’s United Front Work Department paid a marketing company to create videos depicting Uyghurs supporting the PRC government, which a network of inauthentic accounts then amplified on Twitter and YouTube.
中国将其部分外语信息业务外包和私有化，以利用私营部门的创新。中国政府与至少90家中国公司合作，设计面向外国的信息操纵活动，以正面描绘中国。例如，一家由新疆维吾尔自治区广播电影电视局运营并隶属于中共统战部的出版机构向一家营销公司支付费用，要求其制作描绘维吾尔人支持中国政府的视频，然后一个不真实的账户网络在Twitter和YouTube上放大。
Inauthentic Networks Used To Amplify PRC Narratives
Inauthentic networks of bots as well as real accounts that tweet and retweet PRC-approved narratives flood the information space and support astroturfing campaigns. One network of accounts posts information denying atrocities in Xinjiang or accusing “the West” of hypocrisy and another, larger network of accounts amplifies it through retweets and reposting. Stanford University’s Internet Observatory Cyber Policy Center assesses that the PRC’s English-language inauthentic networks have not been successful at gaining traction among foreign audiences.
Influencers Used To Better Reach Young International Audiences
PRC authorities believe social media influencers can help to push PRC messaging to shape local information environments due to their relatability and authenticity. CCP planners seek to adapt how they reach younger media consumers globally and are designing foreign propaganda to be more “youthful” and viral while strictly adhering to political “red lines .” In June 2021, Shen Haixiong, the head of state-run China Media Group – which falls under the direction of the CCP’s Propaganda Department – promoted the use of “multilingual internet celebrity studios ” to enhance the PRC’s image in key regions. Analytics firm Miburo Solutions identified more than 200 third-country influencers affiliated with PRC state media creating social media content in at least 38 languages, including English, Spanish, French, Arabic, and Russian with an average reach of 309,000 followers. Miburo found that the PRC uses influencers to advance its narratives regarding Xinjiang by obscuring state media employees’ affiliations and by orchestrating pro-PRC Western influencers’ tours of Xinjiang.
Trolls Used To Defend PRC Positions and Attack, Insult, and Harass Critics
Internet trolls mainly working under the auspices of the People’s Liberation Army, the Central Cyberspace Affairs Commission, or the Communist Youth League directly attack critics online. According to the French Military School Strategic Research Institute , PRC trolls’ tactics include defending the PRC, attacking and trying to discredit critics, feeding controversies, insulting, and harassing. The PRC’s Cyberspace Affairs Commission and Central Propaganda Department directly employ an estimated two million people nationwide in this capacity and another 20 million working as part-time “network civilization volunteers .” These forces target the PRC’s domestic audience and Chinese-speaking diaspora communities. In response to the Hong Kong protests in 2019, the PRC started to invest more in influencing users of U.S.-based platforms, such as Twitter, Facebook, and YouTube, as well as international platforms, such as VKontakte and Telegram. In 2021, cybersecurity firm FireEye’s Mandiant Threat Intelligence arm and Google’s Threat Analysis Group identified elements of an ongoing PRC-backed information operation that targeted a range of issues, including Xinjiang, in various languages across 30 social media platforms and 40 websites.
