HRIC Digital Rights Report: May 2026
Welcome to the Digital Rights Report, HRIC’s monthly roundup of what you may have missed in the China digital rights space. We cover cybersecurity, AI, digital transnational repression, the Great Firewall, and much more. Thank you for reading!
This Month: Digital Surveillance and Transnational Repression
In May, new reporting emerged on tools being developed by Chinese private companies for government surveillance, data collection, and analysis; creating in-depth profiles of individuals; and even predicting future behavior. Reports indicate that the scope of these tools is likely to extend beyond China’s borders, as the state and its agents continue to employ digital methods of transnational repression. In addition to surveillance, AI-powered harassment of human rights defenders and dissidents has seen a noticeable increase in recent months. For example, AI deepfakes have been used extensively to harass human rights defenders and activists, especially women, with inappropriate pictures and videos.
China Introduces a Program to Track Foreigners’ Every Move
A leak last month exposed a surveillance program called the “Dynamic Management Platform for Overseas Personnel,” a remote data collection and analysis tool for mass tracking of foreign nationals which was developed for the Public Security Bureau in Zhangjiakou, not far from Beijing. Although the site appears to be a “demo,” it contains real information of people who have lived, or are currently living, in China. This includes a tab that lists foreign journalists who were in Beijing around 2021, with ID pictures that appear to be from official visa applications. Cybersecurity researcher NetAskari, who uncovered the dashboard, told Deutsche Welle: “When you work as a journalist in China, you basically assume you are always on their radar. But what surprised me was simply how easy it was to access this highly sensitive system.”
The Telegraph’s former China correspondent, Sophia Yan, released a public video about her experience being included in the database. She explained that the platform contains her passport number, mobile phone number, and workplace, and that she was labeled “trackable.” The platform contains meticulous details of her movements within China: the 78 times she was recorded at a certain intersection, records of her entries and exits at supermarkets and subway stations, and even a relationship graph used to identify her associates.
The Dynamic Management Platform also contains a web-based dashboard that may be connected to the government’s Xueliang (雪亮 or Sharp Eyes) project. The Sharp Eyes project integrates surveillance cameras installed in rural areas with cameras from private residences, including televisions and smartphones, to essentially use private citizens as the eyes of the state. In a similar vein, the Dynamic Management Platform includes data not only from official surveillance devices, but also privately owned devices such as ski lift cameras.
While currently focused on rural areas, the end goal of the Sharp Eyes project is to become a nationwide surveillance tool, with elements such as “real time tracking and predictive policing.” But, there is evidence that it could easily expand even further. For instance, Vanderbilt University recently uncovered documents revealing that the Chinese firm, Geedge Networks, which already sells a commercial version of the Great Firewall, is developing AI tools that analyze citizens’ internet activity, location data, and telecommunications records. Doing so generates profiles that estimate a person’s so-called “political risk” even before they have taken any public action.
International Attention on the Threat of China’s Digital Surveillance
Meanwhile, a new report by the UN Special Rapporteur on Freedom of Assembly and Association on digital tracking and its effect on human rights describes China as an extreme case of surveillance ubiquity. The mobile platforms WeChat and Alipay generate comprehensive digital footprints: one activist stated that “we know that everyone is followed. Everyone, they have data on everyone.” The report characterizes Tibet as “a giant prison” under 24/7 digital monitoring, and documents how simply performing “day-to-day tasks” leaves protesters in Hong Kong vulnerable to police tracking. The report also identifies China as a source of active transnational repression, particularly in relation to Tibetan human rights defenders, who reported state-linked cellular network and SIM-cloning attacks. The UN Report does provide suggestions for responding to these developments in the age of digital surveillance: it encourages human rights law analysis to incorporate compounding harms; it calls for interdisciplinary expertise to be included; it suggests establishing an assessment “focused on evaluating States’ overall surveillance activities”; and it urges a precautionary approach, arguing that even uncertain risks should be taken seriously to prevent them from becoming overwhelming once it’s too late
In another alarming example of digital overreach, a security researcher successfully hijacked every single Yarbo (雅博) robotic mower, which are sold all around the world, from thousands of miles away. He gained access to cameras, sensors, and spinning blades, and he was able to retrieve live camera feeds, exact GPS coordinates of owners’ homes, and even their Wi-Fi passwords. The test served to demonstrate the potential scale of vulnerabilities within the “Internet of Things,” a vast network of “smart” devices with varying levels of data collection, internet, and/or Bluetooth capabilities.
Transnational Harassment
Female activists who challenge the Chinese Communist Party have been systematically targeted by state-linked coordinated campaigns attempting to destroy their reputations and silence them through gendered shame. While gendered violence and digital harassment of female activists is unfortunately nothing new, the latest strategy of using AI-generated deepfake pornographic images is a significant escalation. Now, some activists are pushing back.
After Apple Peiqing Ni, founder of the UK-based China Dissent Network, posted about Tiananmen commemoration activities, she was tagged repeatedly in bot posts on X containing AI-generated deepfake images portraying her as a promiscuous drug addict. Although Ni reported the harassment to the police, she was told the only option was to report it to X’s moderation system. Yet, X’s automated systems ruled that the posts did not violate its rules, despite its policies against harassment. Ni chose to speak up about the issue rather than remain silent, and the account was eventually suspended after the Guardian contacted X for comment.
Safeguard Defenders’ China Program Director Laura Harth publicly addressed the AI-generated pornographic deepfakes targeting her and named the culprit behind them: the Spamouflage foreign influence network linked to China’s public security system. Harth writes: “We have the attributions. We have had them for years. What we need now is concrete action and accountability. None of this requires more evidence. All it requires is political will.”
In Case You Missed It: Other Digital Rights and Cybersecurity News
Great Firewall, Censorship, and Information Control
China Expands Travel Curbs to Top AI Talent at Private Firms: Chinese AI start-up founders, researchers and executives are now subject to state-imposed travel restrictions, meaning they must seek permission from the authorities in order to travel out of the country. While China frequently constrains the movement of government officials and “sensitive” individuals, the move signals a new level of control over private industry.
CDT’s “404 Deleted Content Archive” Summary for April 2026: April’s most heavily censored topics included backlash against a Ministry of State Security article blaming “foreign organizations” for promoting the “lying flat” lifestyle movement, with online commenters challenging the framing and pointing to domestic economic causes, and many of those responses subsequently deleted, as well as suppression of coverage of an April knife attack in Shenyang that killed multiple people and received no official statement.
HRIC on Twitter/X: The Associated Press’s Pulitzer Prize for International Reporting this year was awarded to an in-depth investigation into China’s surveillance system. The investigation, which spanned nearly three years, revealed that multiple Silicon Valley tech companies, with the knowledge of the U.S. government, were deeply involved in the design and construction of China’s surveillance infrastructure. This system has been used for large-scale population tracking and is directly linked to well-documented cases of human rights abuses.
Netizen Voices on MSS Claim That Foreign Forces Are Funding Chinese Slackers: “If Everyone Slacked Off, Who’d Be Left To Exploit?”: China’s Ministry of State Security recently claimed that unnamed foreign organizations have been funding influencers to spread “lying flat” ideology among Chinese youth, framing the slacker trend as ideological warfare. The claim backfired on social media, with netizens responding with sarcasm and pointed to real grievances like youth unemployment and overwork rather than foreign manipulation.
Translations: On the Coal Mine Explosion in Liushenyu, Shanxi, That Killed 82 Miners and Injured 128: China’s deadliest mining accident in over a decade triggered both public anger and online censorship, with investigations revealing widespread safety violations including fake rectifications, fake shutdowns, falsified blueprints, and the authorities’ suspicious inability to account for 123 of the 247 workers underground at the time.
Netizen Voices: “Money Can’t Get Out, and Neither Can People”: New regulations punishing cross-border brokerage apps for giving mainland retail investors access to overseas stocks without licenses, combined with a ban on overseas travel for key employees of domestic AI firms, prompted widespread online mockery summarized in the phrase “money can’t get out, and neither can people.”
Translation: The Decade-long Death of a Campus Media Outlet, Part 2: The Decline of Student Journalism and the Rise of AI: Part two of an essay on the deletion of Beijing Normal University’s student publication Jingshi Xueren describes the emotional aftermath of the erasure, the outlet’s legacy as a formative experience for its contributors, and reflects on the stakes of preserving authentically human perspectives amid the accelerating displacement of student journalism by AI-generated content.
The Guardian view on the splinternet: where China led, Iran and others are eagerly following: This Guardian editorial argues that China, whose Great Firewall has effectively turned 1.125 billion internet users into the world’s largest intranet, is not an anomaly but a pioneer. Its model of “cyber sovereignty” is now being actively adopted by Iran and other authoritarian states to fragment the global internet into a repressive “splinternet.”
AI Propaganda and Social Media Manipulation
HRIC on Twitter/X: A team from institutions including New York University, Princeton, and the University of California, San Diego, report that the outputs of global mainstream LLMs are already being influenced by government control over media worldwide, with Chinese users experiencing particularly pronounced effects.
The CCP is using AI platforms to further digitize and intelligentize its control over Tibet: DeepZang, a Tibetan-language AI platform launched by China this year, provides Tibetan-language services and outputs content that aligns with the official stance on sensitive issues such as Tibet, religion, and governance.
Silence for Sale: A recent state media report exposed a social media extortion scheme in which operators of an account sent “negative information verification letters” to over 180 companies pre-IPO, threatening damaging coverage unless paid off, the latest iteration of a longstanding Chinese practice known as “news extortion” where Party controls have made the power to publish or suppress information a transactional commodity, with no role for the public interest.
China-linked disinformation campaign fails to disrupt Tibetan elections: Beijing’s long-running “Spamouflage” influence network deployed 90 Facebook accounts and 13 Instagram profiles to attack re-elected Tibetan leader Penpa Tsering and cast doubt on the legitimacy of the 2026 exile parliamentary elections, which were held across 27 countries. Despite using AI-generated imagery and coordinated messaging, the campaign drew virtually no organic engagement and failed to meaningfully reach Tibetan diaspora communities.
Cybersecurity & Data Privacy
Eyes in the sky: Hong Kong police tackle petty crime with drone patrols, but privacy concerns linger: Hong Kong police have rapidly expanded automated drone patrols citywide like tracking visa overstayers, gamblers, and “evasive” individuals, raising unresolved privacy questions, as the force plans to add AI capabilities while declining to specify whether facial recognition will be applied to drone footage.
Chinese authorities arrest 97 Tibetans in the name of “Cybersecurity Law”: In 2025, Chinese police arrested 97 Tibetans under the Cybersecurity Law for livestreaming and posting videos in Tibetan on platforms such as Douyin, with authorities setting up 27 special task forces conducting cross-regional investigations across more than 40 regions to crack down on Tibetan-language content. The enforcement has had a severe chilling effect, with many Tibetans stopping all online activity out of fear.
‘Clearly me’: Chinese AI drama accused of stealing faces: A Chinese model and an influencer discovered their likenesses had been used without consent in an AI-generated microdrama on ByteDance’s Hongguo platform, depicting them in unflattering villain roles. The case highlights legal grey areas around AI-generated content in China, with the platform having already taken down 670 non-compliant AI microdramas, and one victim planning to sue.
American staffers told to ditch burner phones before leaving Beijing over fears of Chinese ‘spying’: The security measures underscored a stark contradiction at the heart of the summit: publicly cordial engagement with Xi, while privately operating under the assumption that every device and interaction could be compromised.
Digital Transnational Repression
More private health records of UK Biobank volunteers appear on Chinese website: Health data from all 500,000 UK Biobank volunteers was listed for sale on Alibaba in at least three separate listings, traced to three Chinese research institutions that had downloaded the data under legitimate contracts; the institutions have since had their access revoked.
Civil Society & Resistance Under Pressure
Chinese court awards compensation to sacked worker replaced by AI: A Hangzhou appeals court ruled that a fintech company acted illegally when it demoted and dismissed a quality-assurance supervisor after his role was taken over by AI, ordering over 260,000 yuan in compensation on the grounds that AI cost-efficiency does not constitute the “objective change in circumstances” required by Chinese labor law to justify termination.
Netizen Voices: Is Xi Jinping on Top of the World, or a Tortoise on a Utility Pole?: Even as Xi Jinping hosted high-profile summits with Trump and Putin, a viral Douyin meme framed him as a tortoise perched on a utility pole, seemingly elevated but actually precarious, prompting online speculation about elite political tensions, succession, and the fragility of his position.
Demanding Humanity in Propaganda: When Hunan Economic Television debuted AI-generated news anchors on May 4, public outrage at the prospect of AI-generated content prompted the broadcaster to add a disclaimer clarifying that while the anchors’ images were AI-generated, all reporting and editing remained in human hands, an irony noted widely, given that those human hands remain under strict Party propaganda controls.
The Political Limits of China’s AI Diffusion Ambitions: China’s ambitions to rapidly diffuse AI through its economy are constrained by concerns over employment stability, with youth unemployment at nearly 17% as of March 2026. Chinese courts have begun ruling that companies cannot fire workers simply to replace them with AI, a development that creates a practical political brake on large-scale automation.
Technological Competition & State Power
HRIC on Twitter/X: Anthropic recently released a report on AI that reveals the core strategic logic of the current global artificial intelligence competition and totalitarian AI. This is not merely a contest over technology and commerce, but a critical watershed that will determine the future model of global governance and the baseline for basic human rights.
The Chinese lesson on the human rights approach to AI: A human rights approach to AI must aim at rebalancing power between corporate-state actors, machines, and ordinary people through stronger participatory decision-making frameworks. China’s experience may serve as a cautionary lesson, writes Oiwan Lam.
The rapid embrace of AI in China, its biggest testing ground, may shape how AI is used globally: More than a year after DeepSeek stunned the world, China has become a mass testing ground for AI tools, with Chinese people and businesses rapidly integrating the technology into nearly every aspect of daily life and work. With over 600 million generative AI users as of late 2025, China’s scale of adoption may set global precedents for how AI is used in consumer and enterprise contexts.
International Tibet support groups report: The CCP is promoting its “Tibetan stability maintenance model” globally: A new International Campaign for Tibet report documents how China has used Tibet as a “foundational laboratory” for developing a sophisticated system of high-tech repression, including grid-based surveillance, preemptive policing, and mass DNA databases, that it is now packaging and exporting to foreign security forces through training programs linked to Xi’s Global Security Initiative.
China tech giant Huawei touts new chipmaking technology to sidestep US restrictions: Huawei’s semiconductor chief announced a new chipmaking approach called the “Tau Scaling Law” that optimizes chip communication time rather than transistor density, which the company claims will allow it to produce chips equivalent to next-generation 1.4nm ones by 2031 without requiring the advanced EUV lithography machines it has been denied under US export controls.