U.S. companies built China's surveillance state | Press review n°23
I selected important news on artificial intelligence during the week of September 8 to 14, 2025. Here's my commentary.
Welcome to the twenty-third press review of Artificial reality. This week I focused on the role of U.S. companies in China’s mass surveillance, on the mysterious death of a former OpenAI employee and on the interest of a major Silicon Valley figure for the Antichrist. Have a good read!
📰 Read
U.S. companies built China's surveillance state
Many tech companies from the United States designed and built much of China’s surveillance state over the past twenty‑five years, according to an investigation by the Associated Press.
The Middle Kingdom is known for having implemented a mass surveillance of its own citizens. It is, however, surprising to learn that this surveillance infrastructure was to a large degree put in place by companies based in the United States, a country that claims to defend freedoms and democracy worldwide.
U.S. companies “sold billions of dollars of technology to the Chinese police, government and surveillance companies, despite repeated warnings from the U.S. Congress and in the media that such tools were being used to quash dissent, persecute religious sects and target minorities,” the Associated Press reports. In doing so, they played “a far larger role in enabling human rights abuses than previously known,” the news agency adds.
Predictive policing
U.S. surveillance technologies have allowed, among other measures, “a brutal mass detention campaign” in the far west region of Xinjiang, where virtually the entire native Uyghur population, a Muslim minority, has been targeted and tracked by authorities “to forcefully assimilate and subdue them.”
This was made possible by predictive policing technology exported to China by these U.S. companies. This surveillance method aims to prevent crime, protests, or terror attacks before they happen. To do so, computer systems capture and analyze vast amounts of data —texts, calls, payments, flights, video, DNA swabs, mail deliveries, the internet —to identify individuals deemed suspicious and predict their behavior. “They also allow Chinese police to threaten friends and family and preemptively detain people for crimes they have not even committed,” the Associated Press notes.
For example, the news agency discovered that the Chinese defense contractor Huadi worked with IBM to develop a major policing system called the “Golden Shield.” It enables Beijing to censor the internet and monitor hundreds of thousands of people online in order to crack down on alleged terrorists and “troublesome” individuals.
An accounting ledger leaked by a whistleblower shows that Chinese police later spent tens of millions of dollars acquiring products from IBM, Cisco, Oracle, and Microsoft to upgrade that system.
“Everything was built on American technology”
Nvidia, IBM, Dell, Cisco, Seagate, Intel, Oracle, Motorola, Amazon Web Services, Microsoft, and Western Digital all deliberately sold technology or services to Chinese police or surveillance firms in the country, the Associated Press reveals. Their sales pitches sometimes even quoted Communist Party catchphrases on crushing protests, such as “stability maintenance” and “abnormal gatherings.” Some of their programs for stifling dissent were named “Internet Police” and “Sharp Eyes.”
The Associated Press provides additional examples of controversial contracts:
In 2011, IBM acquired i2, a police surveillance program that analyzes social networks, among others. IBM then hired a Shanghai‑based firm to sell that software to Chinese police, including in Xinjiang.
Nvidia and Intel partnered with China’s three largest surveillance firms to add artificial intelligence capabilities to cameras deployed across the country, including in Xinjiang and Tibet, until sanctions were imposed. Nvidia said in a post dating to at least 2013 that it collaborated with a Chinese police research institute on advanced surveillance technologies.
IBM, Oracle, HP, and Esri sold hundreds of thousands of dollars’ worth of geographic and mapping software to Chinese police that allows officers to detect when blacklisted Uyghurs, Tibetans or dissidents move away from certain provinces or villages.
Dell sold cloud software and storage devices to police and entities providing data to police in Tibet and Xinjiang, even in 2022 after human rights violations there became widely known.
IBM worked with Huadi to build China’s first national fingerprint database. HP and VMware sold fingerprint comparison technology used by Chinese police, while Intel partnered with a Chinese fingerprinting company to make their devices more effective.
IBM, Dell, and VMware promoted facial recognition to Chinese police.
China’s police and police DNA labs purchased software and equipment from Dell and Microsoft to save genetic data on police databases.
In 2016, Dell stated on its WeChat account that its services assisted the Chinese internet police in “cracking down on rumormongers.”
In 2022, Seagate posted on WeChat that it sold hard drives “tailor made” for AI video systems in China for use by police to help them “control key persons.”
“Everything was built on American technology,” says Valentin Weber, a researcher at the German Council on Foreign Relations who studied the use of U.S. tech by Chinese police. “China’s capability was close to zero.”
The foundations of China’s surveillance state
Export of U.S. technologies to China slowed considerably from 2019, following the outrage and sanctions over abuses in Xinjiang. Nevertheless, that “flood of American technology” laid the foundation for China’s surveillance apparatus that Chinese companies have since built on and in some case replaced, the Associated Press notes.
Watch: How US Tech Enabled China’s Surveillance State (Associated Press)
Experts on international sanctions point out that the laws contain significant loopholes and often lag behind new developments. For instance, a ban on selling military and police equipment to China after the 1989 Tiananmen Square massacre does not cover newer technologies or general‑use products that can also be applied in policing.
Concerns therefore persist about how U.S. technologies are used in China. In late July, twenty former U.S. officials and national security experts publicly criticized a deal allowing Nvidia to sell its H20 chips —used in AI systems— to China, with 15 % of the revenue going to the U.S. government. They argue that no matter who the chips are sold to, they will end up in the hands of Chinese military and intelligence services.
Nvidia said it does not make surveillance systems, does not work with police in China, and did not develop the H20 chips for police surveillance. Yet in 2022 the company posted on WeChat that Chinese surveillance firms Watrix and GEOAI were using its chips to train AI patrol drones and systems to identify people by their walk. Those commercial relationships are no longer active, according to the multinational.
Many other countries involved
The Associated Press points out that German, Japanese, and Korean firms have also collaborated with the Chinese government for the surveillance of its population.
In September 2020, Amnesty International warned that several European companies had equipped China with surveillance technologies, including facial recognition systems, digital surveillance cameras and devices that analyze emotions, behavior, gender, age, and ethnicity of the persons filmed, Le Monde reported.
Seven important news this week
ICE Spends Millions on Clearview AI Facial Recognition to Find People ‘Assaulting’ Officers (404 Media)
How ICE Is Using Fake Cell Towers To Spy On People’s Phones (Forbes)
Tech worker dissent over Gaza bubbles inside Amazon, Microsoft and Google (The Washington Post)
Meta suppressed research on child safety, employees say (The Washington Post)
FTC orders AI companies to hand over info about chatbots’ impact on kids (The Verge)
The MechaHitler defense contract is raising red flags (The Verge)
Read the other articles of the week I have selected by clicking here.
🎥 Watch
The mysterious death of a former OpenAI employee
Conservative political commentator Tucker Carlson interviewed Sam Altman, the CEO of OpenAI, in a video published last Wednesday. They talk about different topics such as the suicide of a teenager who had asked ChatGPT about the possible ways of taking his own life.
They also address the controversial death of a former OpenAI employee who was found dead in his apartment in San Francisco a few months after he left the company. Suchir Balaji claimed that OpenAI is not respecting copyright law and that ChatGPT and other chatbots are destroying the commercial viability of the individuals, businesses and internet services that created the digital data used to train these AI systems.
The police concluded to a suicide but Suchir Balaji’s parents, who saw signs of an apparent struggle in their son’s apartment, among other things, think he has been killed.
Watch: "Signs of Struggle": Parents Seek Truth On OpenAI Whistleblower Death (Breaking Points) and OpenAI Whistleblower Found Dead: Suicide or Murder? (Breaking Points)
Tucker Carlson presses Sam Altman on this subject, and even tells him that Suchir Balaji’s mother claims her son was killed “on his orders”. This passage starts at 34:28.
Sam Altman on God, Elon Musk and the Mysterious Death of His Former Employee (Tucker Carlson)
Seven important videos this week
ICE reactivation of spyware contract raises alarm about next steps (MSNBC)
'Full Of Lies', Suchir Balaji's Mother Exposes Sam Altman's Lies | Poornima Rao (Republic World)
The ChatGPT Messages OpenAI Doesn't Want You to See (Vanessa Wingårdh)
Superhuman AI: If Anyone Builds It ... Everyone Dies? (Semafor)
Big Tech Under Pressure: Hunger Strikes and the Fight for AI Safety (The AI Risk Network)
How AI Companies Created a Fake Arms Race (Species | Documenting AGI)
Demis Hassabis: The CEO Working to Solve Cancer With AI (Bloomberg)
Watch the other videos of the week I have selected by clicking here.
🔈 Listen
Peter Thiel and the Antichrist
In the new episode of The Nerd Reich, journalist Gil Duran comments on the new lecture series of Peter Thiel, the cofounder of data analysis company Palantir, an ally of Trump and one of Silicon Valley’s most influential figures. These conferences started on Monday in San Francisco. The topic? The Antichrist.
Watch: Peter Thiel and the Antichrist (Interesting Times with Ross Douthat)
Gil Duran interviews Robert Fuller, author of «Naming the Antichrist: The History of An American Obsession», and theologian Matthew Fox. They explain how the myth of the Antichrist has been used in the past to justify extremism and push society towards violence.
Listen to The Nerd Reich on PodBean
Thank you for reading the twenty-third press review of Artificial reality! Subscribe for free to receive the reviews directly into your mailbox. With a paid subscription, you will also have access to all articles and to an exclusive monthly newsletter.
Have a good week,
Arnaud
