Welcome to the HRIC Digital Rights Report, our monthly roundup of developments in China’s digital rights landscape, including cybersecurity, AI, digital transnational repression, censorship, the Great Firewall, and internet freedom.
This Month: Digital Censorship Reaches Beyond China’s Borders
In October, French football star Kylian Mbappé faced significant backlash in China after a video resurfaced in which he was asked about his favorite internet meme. Mbappé referred to an image of his face superimposed onto that of a “Chinese dictator,” an apparent reference to Mao Zedong, although he did not mention Mao by name. Mao, who founded the People’s Republic of China in 1949 and ruled until his death in 1976, remains a politically sensitive figure in China, with his legacy of political repression and policies that resulted in millions of deaths. Following Mbappé’s brief remarks, at least three Chinese brands terminated their commercial relationships with him and the Chinese state broadcaster CCTV suspended broadcasts of Spain’s La Liga football matches. The Chinese Football Association also demanded formal apologies from Mbappé and Real Madrid. On October 27, Mbappé publicly apologized, describing his comment as “ill-judged,” while Real Madrid and La Liga also issued apologies the following day.
On October 4, well-known Taiwanese host Tsai Kangyong (Kevin Tsai) attended the launch event for the campaign headquarters of Shen Boyang, Taiwan’s Democratic Progressive Party’s candidate for Taipei mayor, in preparation for potential interviews with the candidate. After Tsai attended Shen Po-yang’s campaign-headquarters launch, Taiwan’s Central News Agency reported that his Weibo account with 36 million followers had all its content wiped clean and all followers reset to zero.
These incidents serve as a troubling reminder of how far China’s influence can reach beyond its borders, affecting not only Chinese nationals abroad who face transnational repression, but also foreign nationals who are not directly criticizing the Chinese government. Mbappé’s comment was neither a political statement nor a rebuttal of China’s government, but rather a passing reference to an internet meme. Nevertheless, in a short amount of time, it quickly resulted in commercial repercussions and broadcasting restrictions. This will likely create a broader chilling effect, encouraging international celebrities and organizations to avoid even minor remarks about politically sensitive subjects for fear of losing access to the Chinese market.
Alongside these measures, Chinese authorities have continued to actively shape international perceptions through social media and partnerships with foreign influencers in the past month. On October 18, China Digital Times published a translation of a Chinese-language commentary examining government-funded influencer campaigns and questioning why favorable foreign portrayals of China receive relatively little scrutiny. The article reported that Hainan provincial authorities allocated up to 894,000 yuan (approximately $133,000) for a project involving international influencers to promote tourism in the region, including expenses related to travel, accommodation, content production, and publicity. A March 2026 report by the European Council on Foreign Relations also identified a strategy known as “borrowing mouths,” through which Chinese state-affiliated actors use foreign influencers and local intermediaries to promote narratives favorable to Beijing while hiding the government’s involvement. Similarly, a May 2026 investigation by Le Monde highlighted how China-related travel content is increasingly shaping Gen Z’s perceptions of the country, particularly as younger audiences rely on TikTok and Instagram rather than traditional media for information. This is particularly interesting because, while the Great Firewall blocks access to major international social media platforms within China, Chinese government-affiliated organizations actively use these same platforms to manipulate international perceptions of China and promote a favorable image of the country abroad.
(Notably, last month Harbin Engineering University gave 20 students disciplinary warnings for using VPNs to access “overseas websites” and conduct research. The warnings were allegedly given after another student informed on them, sparking commentary on a resurgence of “informer culture” as well as the necessity of using circumvention methods to carry out daily work.)
China-linked digital operations are not limited to shaping international perceptions. They also target diaspora communities living abroad, although these activities often receive less public attention. On October 15, Phayul reported on findings by the Tibetan Computer Emergency Readiness Team (TibCERT), which documented two cyber espionage campaigns targeting Tibetans around the Dalai Lama’s 91st birthday. In one campaign, attackers linked to the China-associated hacking group Mustang Panda sent deceptive emails using politically sensitive Tibetan events to trick recipients into downloading malware that could give attackers access to their computers. In another, attackers compromised a Tibetan community website to collect information about visitors and spread malicious software. TibCERT recorded 49 reported security incidents during the quarter and warned that the actual number was likely higher.
In Case You Missed It: Other Digital Rights and Cybersecurity News
Great Firewall, Censorship, and Information Control
WeChat, QQ Users Banned for Hiding the Emperor’s New Clothes: Users have reported permanent suspensions from WeChat and QQ after posting the sentence “Lately, I’ve grown used to quiet, uneventful days” (Xíguàn le jìnlái píngpíngdàndàn de rìzi), which likely “triggered keyword scans with its non-sequential incorporation of Xi Jinping’s name.” Users complain that platforms tend to simply accept false positives as preferable to missing any politically sensitive content.
Translation: Censored Commentary on Beijing’s Drone Storage Ban: A mild WeChat post about Beijing’s new rule banning drone possession, storage and sale citywide from November 15 was deleted. The post’s removal fits a pattern of censoring even non-critical commentary on security measures.
Hong Kong national security police arrest insurance agent and woman over ‘seditious’ social media pages: In Hong Kong, national security police arrested one man and one woman on October 6 over allegedly seditious social media pages, reportedly the “YellowInsurance” platform that mixed insurance promotion with pro-protest content. Sources identified the man as a licensed insurance agent and ex-localist, and the woman as a friend suspected of helping manage the accounts.
Hong Kong nat. security police charge man over ‘seditious’ posts 3 months after unannounced arrest: Hong Kong charged a 52-year-old man over social media posts allegedly meant to provoke hatred towards the central and Hong Kong governments and the judiciary. The case was publicly announced only after his court appearance, breaking the usual police practice of issuing a statement when charges are filed.
What’s the impact of China’s influence on the gaming industry?: Chinese game companies are increasingly shaping the global industry. Yet, games developed from within China face strict censorship and Chinese gamers must register using their real names, which allows monitoring of in-game communications.
Who Hit the Brakes on Negative Public Opinion?: A Dongchedi test video showing the brake pedal bracket snapping on Huawei-backed Maextro’s V800 minivan disappeared from platforms within 24 hours, prompting discussion online related to who was responsible for removing the video. A week later, the industry ministry opened a safety probe, with an unusual caveat: the “Cyberspace Administration of China (CAC), the country’s top internet control body, would be joining the investigation to look into the problem ‘reflected online.’”
冯正虎:一篇文章”违规”,整个账号被封停?——正虎纪实百家号账号被封事件始末 [Feng Zhenghu: One article “violation” and the entire account was suspended? — The whole story of the suspension of Zhenghu’s Baijiahao account]: The Baijiahao account Zhenghu Jishi was banned on October 18 after one article describing the author’s experience in promoting rule of law and the Constitution was found to “contain violations.” The author appealed, arguing only the single article should be pulled. Due to China’s real-name registration laws, the harshly disproportionate measures taken against accounts that are found to have crossed censorship “red lines” have a significant impact on individuals’ ability to re-access internet platforms.
CDT’s “404 Deleted Content Archive” Summary for September 2026: CDT archived 31 deleted pieces from the Chinese internet covering topics including the Sun Meihua case (a Beijing woman found in Shandong after 26 years), a Guangzhou stabbing which targeted women, the Shaoyang 100-million-yuan official extortion case, Beijing’s drone-storage ban, Xingyu workers’ labor activism, and Wuhan’s fines on Guo Degang’s performance organizers.
AI Propaganda and Social Media Manipulation
HRIC on Twitter/X: Taiwan’s Center for Technology, Democracy, and Society recently released a report examining censorship biases in two mainstream large language models, ChatGPT-5.2 and Gemini 3. The research team queried the models using 400 terms in English, Traditional Chinese, and Simplified Chinese, collecting a total of 2,400 responses. Among them, 300 were politically sensitive terms that could be subject to CCP censorship, while another 100 were natural science vocabulary used as a control. The results showed that the Chinese responses from both models were more evidently aligned with censorship narratives compared to the English responses.
HRIC on Twitter/X: A cybersecurity firm tested Alibaba’s open-source large model Tongyi Qiwen (Qwen) and found it evades the facts on sensitive political issues. Qwen has been downloaded globally over 3 billion times. It is also used by major multinational companies like Airbnb and Uber.
China accelerates DeepZang AI upgrades under ‘ethnic integration’ agenda: At an October 5 seminar in Hohhot, the developers of DeepZang, a Tibetan-language LLM launched in March, discussed better dialect handling, real-time Tibetan-Mandarin translation, and wider use in education, healthcare and governance. An ICT report found the model repeats official Chinese narratives on sensitive topics and restricts others, such as self-immolation protests.
China’s AI Regulation in 2026: New Laws, Rules, and Compliance Requirements. A comprehensive AI law is on the 2026 legislative agenda, but no draft exists yet. The amended Cybersecurity Law added an AI provision, rules on human-like companion AI took effect July 15, and an AI agent policy and the September AI Safety Governance Framework 3.0 extend regulation to agents.
2026 AI Regulations in China: Enforcement, Guidance and Compliance: Regulators moved to active enforcement in April, acting against Jianying, Maoxiang and Jimeng for failing to implement AI-content labelling rules. The Supreme People’s Court’s September 7 opinions set notice-and-takedown duties for AI providers and address voice cloning and training data.
Opinions on Adapting to the Development of Artificial Intelligence Technology to Help Build an Employment-Friendly Development Model: The October 23 guideline sets 18 tasks in six areas to grow AI jobs while cushioning displacement. It urges “unemployment buffer periods” and advance notice to affected workers, employment-risk assessments for large AI projects, and a cautious pace in labor-intensive sectors.
AI Authorship in a Party Mouthpiece: Jilin Daily ran an autumn essay bylined “AI” on October 12, which media circles called a first for a Party paper. CMP says the piece was part of an essay contest inviting readers to compete with AI, although the paper did not disclose the model or provide readers with that context.
Taiwan says China cyberattacks, disinformation up ahead of local elections: Taiwan’s National Security Bureau says China spread over 2.5 million posts containing disinformation, primarily through Facebook and Threads, in the first nine months of 2026, about 48% more than the year prior. It warns of rising operations ahead of November 28 local elections, focused on energy, food security, defense and candidates. Beijing has denied any such interference.
Graphika finds China-linked Spamouflage accounts posing as US voters on X: Graphika found three China-linked account sets on X posing as American voters and posting divisive content on politics, social issues and the midterms. Two of the sets had earlier been typical Spamouflage accounts that attacked Guo Wengui and Falun Gong before switching to US-focused content.
Cybersecurity & Data Privacy
USALI This Week in Asian Law: September 27 – October 03: Beginning October 1, 2026, the Ministry of Public Security has expanded powers to inspect companies’ handling of data and personal information. The new Measures for Public Security Organs’ Supervision and Inspection of Cyberspace Security update 2018 rules to reflect the 2021 Data Security Law and Personal Information Protection Law.
China moves to curb tech exports, including AI, citing national security risks: China’s commerce ministry proposed restricting or banning exports of 40 technologies spanning AI, chips and solar, citing national security. Exports of AI models trained on sensitive national data and defense-focused AI agents would be banned, and models with 2 trillion or more parameters could be restricted.
DeepSeek barred from government agency IT systems: The U.S. Department of Homeland Security ordered civilian federal agencies to strip DeepSeek products from their IT systems and stop buying them, including third-party tools built on its models. The order went out quietly in July and was published October 19 under the 2018 Federal Acquisition Supply Chain Security Act, apparently its first use against an AI company. DHS cited risks including collection of personal data, user fingerprinting, possible malware, and “transfer of sensitive data to the PRC.”
Digital Transnational Repression
How Generative AI is Scaling Transnational Repression: Citizen Lab researchers say Anthropic’s September threat report is concrete evidence that AI is being built into cross-border repression, citing a China-aligned operator who used Claude to track Uyghurs in Syria and recruit informants. They urge AI companies to publish dedicated reports, improve detection and refusals, notify affected communities and preserve evidence.
Technological Competition & State Power
MI5 warns UK universities over ‘theft of tech secrets by Chinese front company’: In its first public “espionage alert,” the UK’s MI5 stated that the China General Technology Research Institute is a front tied to China’s Ministry of State Security, and that the Institute had already funded projects involving 100+ UK-linked academics in areas including AI and cybersecurity.
HRIC on Twitter/X: U.S. cybersecurity firm Proofpoint disclosed that a hacker group suspected to be linked to the Chinese government impersonated prominent figures in the U.S. AI sector in February and July this year, launching phishing attacks against experts researching AI export controls and military AI applications at U.S. universities, think tanks, and law firms, with the intent to steal their email accounts and gain insight into how U.S. society discusses and formulates AI policies.
US arrests California man for allegedly smuggling $300m worth of computer servers to China: Greg Lui, owner of Earthmade Computer, was arrested on charges of using false paperwork and shipments through Malaysia and Singapore to route export-controlled GPU servers to China in 2023–24. He faces counts of conspiracy, smuggling and money-laundering, in a case that follows Taiwan’s recent prosecutions over smuggled Nvidia servers.
West-to-east energy, east-to-west computing: The Uyghur costs of China’s digital order: The author argues that Beijing is turning the Uyghur region into a Eurasian AI and data hub, pointing to the SCO Digital Economy Forum in Urumqi, as an extension of its long-standing resource extraction.
The FBI has seized tools used by Chinese hackers for cyber operations, officials say: The FBI and Justice Department seized the scanning tool “Microscan” and phishing tool “FishHub,” used by the Flax Typhoon group, which operated through Chinese firm Integrity Technology Group. They had been used against a US power company, Japanese and Polish airports, and Taiwanese universities and critical infrastructure.
Alibaba sues Pentagon to denounce being classified as a Chinese military affiliate: Alibaba sued over its addition to the Pentagon’s Section 1260H list of Chinese military companies in June, saying the listing has “no basis in fact or law.” The Pentagon’s filing cites Alibaba’s censorship tools, joint ventures with state firms such as Norinco-linked Qianxun, and golden-share arrangements. Alibaba says Chinese-law compliance is not military affiliation.
HRIC on Twitter/X: The U.S. Department of State has offered a bounty for information on the identity and whereabouts of Chinese national Zhang Yu, a director of Shanghai Shiyan Technology Co., Ltd., who reportedly conducted hacking activities on behalf of the Shanghai State Security Bureau of China’s Ministry of State Security.
‘Super app’ aimed at UK students uses servers owned by Chinese firm on US blacklist: “ZyMix, which is being pushed on university campuses as the UK’s first super app designed ‘for gen Z’, combining chat, payments and community features, has a contract with a German-based subsidiary of Tencent Holdings,” the Guardian reports. Tencent was recently put on a list of firms that the U.S. government alleges work closely with the Chinese military.
China-linked hackers hit Fiji: US embassy: China-linked hackers reportedly kept access to the network of an unnamed Suva-based regional organization for several days between July and August. The report frames the breach as part of wider targeting of Pacific governments and telecoms, while China’s embassy rejected the claim as evidence-free politicization.
Warning Signals: Venezuela and the Risk of AI-Powered Chinese Digital Authoritarianism: Venezuela, which already runs a surveillance system built largely on Chinese Huawei, ZTE and Hikvision equipment, could add Chinese AI through a July 2025 memorandum with sanctioned firm iFlytek. Facial and voice recognition combined with the Carnet de la Patria ID database could enable preemptive profiling.

