Facebook’s AI treats Palestinian activists like it treats American Black activists. It blocks them.
Elizabeth Dwoskin, Gerrit De Vynck3 days ago
Liberal Media Scream: ‘Some things may be true even if Donald Trump said them’
100 bullets fired in mass shooting that left 2 dead, 20 hurt at birthday bash: Police
Just days after violent conflict erupted in Israel and the Palestinian territories, both Facebook and Twitter copped to major faux pas: The companies had wrongly blocked or restricted millions of mostly pro-Palestinian posts and accounts related to the crisis.
Activists around the world charged the companies with failing a critical test: whether their services would enable the world to watch an important global event unfold unfettered through the eyes of those affected.
The companies blamed the errors on glitches in artificial intelligence software.
In Twitter’s case, the company said its service mistakenly identified the rapid-firing tweeting during the confrontations as spam, resulting in hundreds of accounts being temporarily locked and the tweets not showing up when searched for. Facebook-owned Instagram gave several explanations for its problems, including a software bug that temporarily blocked video-sharing and saying its hate speech detection software misidentified a key hashtag as associated with a terrorist group.
The companies said the problems were quickly resolved and the accounts restored. But some activists say many posts are still being censored. Experts in free speech and technology said that’s because the issues are connected to a broader problem: overzealous software algorithms that are designed to protect but end up wrongly penalizing marginalized groups that rely on social media to build support. Black Americans, for example, have complained for years that posts discussing race are incorrectly flagged as problematic by AI software on a routine basis, with little recourse for those affected.
Despite years of investment, many of the automated systems built by social media companies to stop spam, disinformation and terrorism are still not sophisticated enough to detect the difference between desirable forms of expression and harmful ones. They often overcorrect, as in the most recent errors during the Israeli-Palestinian conflict, or they under-enforce, allowing harmful misinformation and violent and hateful language to proliferate, including hoaxes about coronavirus vaccines and violent posts ahead of the U.S. Capitol insurrection on Jan. 6.
The Palestinian situation erupted into a full-blown public relations and internal crisis for Facebook. Last week, CEO Mark Zuckerberg dispatched the company’s top policy executive, Nick Clegg, to meet with Israeli and Palestinian leadership, according to the company. Meanwhile, Palestinians launched a campaign to knock down Facebook’s ranking in app stores by leaving one-star reviews. The incident was designated “severity 1” — the company’s term for a sitewide emergency, according to internal documents reviewed by The Washington Post and first reported by NBC. The documents noted that Facebook executives reached out to Apple, Google, and Microsoft to request that the posts be deleted.
Meanwhile, a group of 30 Facebook employees, some of whom said they had friends and family affected by the conflict, have complained of “over-enforcement” on the Palestinian content in an open letter on the company’s workforce messaging boards, according to another set of internal documents reviewed by The Post. The group has filed at least 80 tickets to report “false positives” with the company’s automation systems in relation to the conflict, noting many of the problems were with the AI mistakenly labeling images of protests as “harassment or bullying.”
Jillian York, a director at the Electronic Frontier Foundation, an advocacy group that opposes government surveillance, has researched tech company practices in the Middle East. She said she doesn’t believe that content moderation — human or algorithmic — can work at scale.
“Ultimately, what we’re seeing here is existing offline repression and inequality being replicated online, and Palestinians are left out of the policy conversation,” York said.
Facebook spokeswoman Dani Lever said the company’s “policies are designed to give everyone a voice while keeping them safe on our apps, and we apply these policies equally.” She added that Facebook has a dedicated team of Arabic and Hebrew speakers closely monitoring the situation on the ground, but declined to say whether any were Palestinian. In an Instagram post May 7, Facebook also gave an account of what it said led to the glitch.
Twitter spokeswoman Katie Rosborough said the enforcement actions were “more severe than intended under our policies” and that the company had reinstated the accounts where appropriate. “Defending and respecting the voices of the people who use our service is one of our core values at Twitter,” she said.
Palestinian activists took to the social media platforms as they began staging protests in late April ahead of an impending Israeli Supreme Court case over whether settlers had the right to evict families from their homes in the Jerusalem neighborhood of Sheikh Jarrah. Potential evictees live-streamed confrontations and documented footage of injuries after Israeli police stormed al-Aqsa Mosque, one of the holiest sites in Islam.
The conflict descended into war after terrorist group Hamas, which governs Gaza, fired explosive rockets into Israel. Israel responded with an 11-day bombing campaign that killed 254 Palestinians, including 66 children. Twelve people in Israel were killed, including two children.
During the barrage, Palestinians posted photos on Twitter showing homes covered in rubble and children’s coffins. A cease-fire took effect May 20.
Palestinian activists and experts who study social movements say it was another watershed historical moment in which social media helped alter the course of events. They compared it to a decade ago, when social media platforms were key to organizing the pro-Democracy uprising known as the Arab Spring. But at the time, tech companies didn’t rely on policing algorithms, rather humans making decisions. And while mistakes were made, nothing occurred on the scale of today, York said.
Even after the companies said the glitches were fixed, 170 Instagram posts and five Twitter posts that activists believe were wrongly removed were still offline, according to 7amleh, the Arab Center for the Advancement of Social Media, a group that advocates for Palestinian digital rights. The group said in a report in late May that it was told by the companies that some of the remaining posts are under review.
Facebook declined to comment. Twitter’s Rosborough said she could not comment without seeing the tweets.
During the early protests in East Jerusalem, some posts on Facebook, Twitter and Instagram were taken down for using the hashtag #SaveSheikhJarrah, the name of the neighborhood in dispute, said Iyad Alrefaie, director of Sada Social, a group that tracks digital rights in the Palestinian territories.
Mariam Barghouti, a Palestinian-American journalist who covers the West Bank for Al Jazeera and other outlets, posted on Instagram that she had her account restricted by Twitter for purportedly violating the company’s social media policy while covering a protest. She said in media interviews that she did not know which tweets broke the rules. The company later restored her account and tweets, saying it made an error, according to spokeswoman Rosborough.
Digital rights groups Access Now, 7amleh and other organizations have spent the years since the Arab Spring documenting problems with how social media companies handle Palestinian content, as well as content from the region at large.
In 2016, Facebook blocked the accounts of several editors at two Palestinian news organizations without giving a reason, Al Jazeera reported at the time. After complaints, the social media company reversed the bans and said they had been accidental. In 2019, Twitter suspended accounts run by a Palestinian news organization, Quds News Network, in a sweep of terrorist accounts (which have since been reinstated, Twitter said). In May 2020, Facebook deactivated the accounts of more than 50 Palestinian journalists and activists without providing an explanation, activists said, including from journalists who posted footage of attacks by Israeli settlers on Palestinian farmers in occupied territories.
Facebook declined to comment on those examples.
Facebook took down a post from a father wishing his infant son, named Qassam, a happy birthday, according to Alrefaie, the director of Sada Social. The group assumed that it was because the company blocks many posts about al-Qassam Brigades, Hamas’s military wing.
“These words are part of our discourse, it’s a part of our culture,” Alrefaie said. “Facebook didn’t differentiate between any context.” Facebook declined to comment on that incident.
Marwa Fatafta, digital rights policy manager for the Middle East and North Africa region for Access Now, said other keywords, such as the term Zionist, are often banned when Palestinians use them because it’s assumed to be antisemitic.
“Under our current policies, we allow the term ‘Zionist’ in political discourse, but remove attacks against Zionists in specific circumstances, when there’s context to show it’s being used as a proxy for Jews or Israelis, which are protected characteristics under our hate speech policy,” Facebook’s Lever said.
Some activists have developed workarounds to the algorithms, including using an ancient method of writing Arabic, according to an article by independent Egyptian news website Mada Masr. Some U.S. activists use similar tactics, purposely misspelling common words like “white” to avoid algorithmic censorship during discussions of race, The Post has reported.
Activists have also decried tech companies’ relationship with the Israeli government, and in particular the Ministry of Justice’s Cyber Unit — which has a direct channel to technology companies to report potential content violations. They have asked tech companies to be transparent about when the government secretly refers accounts to be blocked or content to be removed, including whether the unit was involved in takedowns during the war.
Facebook, Google and Twitter all said they comply with local laws and regularly respond to takedown requests from governments, which they publish in biannual transparency reports. Twitter said the spam filter issue had nothing to do with Israeli authorities. Facebook did not respond to several requests for comment about the nature of reports by Israeli authorities during the recent crisis. A Google spokesman declined to say whether it received bulk requests from the Cyber Unit.
Journalists and activists have also complained that Google hasn’t updated its maps of Gaza with higher-resolution images, despite a U.S. law limiting the degree of detail in public maps of the area being lifted in 2020. Detailed maps help document the damage from airstrikes.
Google declined to comment on why the Gaza maps have not been updated.
Payment app Venmo also mistakenly suspended transactions of humanitarian aid to Palestinians during the war. The company said it was trying to comply with U.S. sanctions and had resolved the issues.
Tech companies are caught between governments trying to stop unrest or violence and activists advocating for free democratic expression, said James Grimmelmann, a law professor at Cornell Tech.
“So the platforms really have to make deeply political choices,” he said.
The latest issues began May 5, when Instagram started receiving reports that people participating in protests in Colombia could not share video, the company later said in a post in which it apologized for its errors. The next day, similar reports came from people participating in demonstrations in Canada and in East Jerusalem. Executives discovered a glitch in a long-planned update to video-sharing products, called Stories. In its apology, the company noted that the bug had nothing to do with these particular events, and in fact had affected more users in the United States than elsewhere.
Several days later, citizens and activists began reporting their posts about al-Aqsa Mosque, using the hashtag #AlAqsa or its Arabic counterparts, were being restricted. The restrictions were often accompanied by a pop-up that said the term was associated with “violence or dangerous organizations.”
On May 11, a Facebook employee filed a grievance, according to a report by BuzzFeed. Facebook said in response that the name of the mosque was a designated terrorist organization. Facebook later told The Post that the hashtag had been restricted in several ways, including limiting people’s ability to search for it.
After publication, Facebook’s Lever added that human error led to the issue of restricting the al-Aqsa Mosque hashtag.
Palestinian activists and Facebook employees began to protest in the coming days that many posts about the conflict were being taken down automatically.
Around the same time, Twitter began fielding reports that influential accounts tweeting about the conflict were being unexpectedly suspended, the company said, due to AI mistaking posts for spam. The company says it restored the accounts a few hours later.
Twitter spokeswoman Rosborough noted that similar incidents of overly severe enforcement took place during the 2020 presidential debates and during protests against a coup this spring in Myanmar.
And sometimes, she pointed out, algorithms get things right: At one point during the conflict, an algorithm also automatically restricted the Israeli army’s official account. The account was trying to post the same tweet twice, of emergency sirens going off in the southern city of Beersheba, and Twitter blocked it.
“We know it’s repetitive — but that’s the reality for Israelis all over the country,” the tweet said.
NBA star Kyrie Irving’s recent conversion to Islam has brought new attention to the relationship between Muslim athletes and Ramadan, the holiest month on the Islamic calendar, when Muslims abstain from food and drink during the day.
“All praise is due to God, Allah, for this … for me, in terms of my faith and what I believe in, being part of the Muslim community, being committed to Islam,” the Brooklyn Nets guard said at an NBA news conference late last month. Irving said he is also “committed to all races and cultures, religions” and that having an understanding and respect for them is what he wants “as a foundation.”
The statement was the first time the Australian-born Irving had mentioned his conversion to Islam. Irving, 29, won world championships with the U.S. National Basketball Team at the 2016 Olympic Games in Rio de Janeiro and the 2014 FIBA Championship in Spain. The seven-time NBA All-Star said Ramadan offers a different kind of challenge.
“I am taking part in Ramadan with a lot of my Muslim brothers and sisters,” he said in the same news conference. “It’s been an adjustment … being committed to my service to God, Allah, and then continuing on with whatever I’m guided with.
“I’m just happy to be part of my community and doing the right things. So fasting is definitely part of it,” he added.
Irving has largely played with rare form during the month, earning him the nickname “Ramadan Kyrie” from fans on social media.
He isn’t the only Muslim NBA player observing fast this Ramadan. Others include Enes Kanter and Jusuf Nurkic of the Portland Trail Blazers, Tacko Fall and Jaylen Brown of the Boston Celtics, and Hamidou Diallo of the Detroit Pistons.
Ramadan is tied to the Islamic lunar calendar and is likely to fall during every NBA season between now and 2036.
Historically, other Muslims in the NBA have fasted during Ramadan, such as former Los Angeles Laker star Kareem Abdul-Jabbar and former Houston Rockets player Hakeem Olajuwon. Olajuwon, a native of Nigeria, went on to become the first foreign-born player to win the league’s MVP award and famously led his team to an NBA finals victory while fasting during Ramadan. Other NBA players have converted to Islam after leaving the game, notably former NBA player Stephen Jackson, who converted earlier this year.
Muslims are exempted from fasting under certain situations. Among those exempted are children, people with health conditions, the elderly, pregnant or nursing women, and those dealing with hardships of travel. Indeed, the reality of life as a modern athlete has allowed some to argue that it is a situation akin to traveling in which fasting is not required and can be resumed at a later date.
“Performance of repeated anaerobic exercise is impaired, but aerobic power and muscular strength show little change during Ramadan,” wrote Roy Shephard, a Canadian doctor, in a 2013 academic study of the impact of Ramadan on athletes. “Ratings of fatigue are increased, and vigilance and reaction times are impaired, particularly during the afternoon. Medical issues during Ramadan are few.”
For those in combat sports, where schedules are more fluid, athletes can cope with Ramadan in different ways.
“I took the route of mostly training at night,” said undefeated boxer Hasim Rahman Jr., who is currently training for his 11th professional fight, scheduled for later this month in Mexico. “After I open my fast, I digest for an hour before training into the night. Today I trained during the day, and that is something I do on occasion. To me, this proves even more that boxing is a mental test as much as anything else,” Rahman told Religion News Service.
Rahman is the son of former heavyweight champion and Baltimore standout Hasim Rahman. The younger Rahman said boxers don’t usually eat while training but drink water regularly during sparring, often taking sips from a water bottle between rounds. He said that although water is beneficial and recommended, a focused athlete can still train without it.
Another Muslim boxer, Julian Williams, won multiple super welterweight world titles for a fight that took place during Ramadan in 2019. Williams wished viewers of the broadcast “Ramadan Mubarak” during his post-fight interview — perhaps a first for American sports television.
Religious authorities in Muslim-majority countries usually cite the hardships of travel when ruling that athletes participating in the Olympics, international soccer or other sporting events can abstain from fasting during Ramadan. One notable exception is Iran, where in 2010 soccer star Ali Karimi was fired by his team for failure to observe the Ramadan fast.
Ramadan is also an issue for amateur Muslim athletes, who are perhaps not traveling to a different city, much less a different country for their events.
In California’s Santa Clarita Valley, a group of five Muslim couples who form the amateur softball team the Better Halves have continued to play in their coed league during Ramadan. The team is very much a family affair and has included players as young as 18 and as old as 70. League officials have tried to schedule games close to sunset to accommodate the needs of the team.
“The lack of energy is balanced out by the pure concentration you have. You go in knowing you will try to save energy, and in the end, we usually play better and still have energy left,” said team captain Arif Harisolia, “During Ramadan, you are a lot more focused on all aspects of life.”
A cyclist rides past a signage displayed outside Oracle’s building at Zhongguancun Software Park on Aug. 30, 2020, in Beijing.Photo: VCG via Getty Images
BANNERS PRINTED FOR the occasion read, “Build a new type of strategic partnership.” Artfully made cutouts of the two companies’ logos adorned the stage. And the frosting on the massive sheet cake curled into a red “20,” to celebrate two decades of cooperation between Oracle and one of its most important Chinese resellers.
This was the backdrop in 2018, when Oracle executives gathered with management from Digital China, a Beijing-based broker used by foreign tech companies to access the Chinese market. Roger Li, Oracle senior vice president and managing director for China, delivered a speech extolling the deep and sustained cooperation between the two firms, according to an account published on Digital China’s website. His team posed for numerous photos with Digital China representatives, including a carefully staged shot in which they together placed their hands on a cake knife.
Photo: Digital China website
The event showed a deepening relationship between Oracle, an Austin-based multinational best known for its databases, and its longtime partner. Soon after, Oracle named Digital China a global “partner of the year.” At a 2019 Oracle software developer conference in Shenzhen, a keynote speaking slot went to Digital China’s chief technology officer. And Digital China’s board includes a former Oracle China vice president.
But Oracle’s relationship with the Chinese broker is deeply problematic. Oracle has pledged to “uphold and respect human rights for all people” and stated that its partners should “have the same mission and vision” as Oracle itself. Co-founder Larry Ellison once criticized a competitor, Google, for going “into China and facilitat[ing] the Chinese government surveilling their people.” And yet Digital China and its network of subsidiaries are critical purveyors of technology used to build China’s surveillance state. Among the company’s offerings are services for Chinese police and defense entities — as advertised on Oracle’s own website.
How Oracle Sells Repression in China
In February, The Intercept revealed that Oracle employees had marketed the company’s analytics software for use by police and military-linked clients in China, as well as by police and paramilitary forces in Brazil, Mexico, Pakistan, and the United Arab Emirates. Documents found on Oracle’s site indicated that at least two Chinese provincial police departments have used the software to crunch surveillance data. At the time, Oracle spokesperson Jessica Moore said that Oracle had “no known implementations” with Chinese police. She described the presentations, which bear the company’s logo, as hypothetical “pitch decks.”
But a former Oracle senior director contradicted that account. At the 2017 Oracle OpenWorld conference in San Francisco, he presented on a surveillance project by an unnamed provincial police department in China. He told The Intercept that the department had used Oracle software.
And even Moore left open the possibility that brokers might have sold analytics software built with Oracle technology to police departments. “Third parties or systems integrators could develop products on top of our technology,” she said, though she claimed that it was unlikely.
Now, The Intercept has unraveled that network of resellers — and found deep and long-standing relationships between Oracle and brokers that sell surveillance technology to the Chinese government.
Join Our Newsletter
Original reporting. Fearless journalism. Delivered to you.
In addition to Digital China, Oracle’s website names as a partner a subsidiary of CEC, a state-owned conglomerate that was listed last year by the Defense Department for its military ties. The website of that subsidiary, Great Wall Computer Software and Systems, describes extensive work with the Ministry of Public Security and provincial public security departments. The broker has helped with “anti-terrorism” work (a term generally used to connote the crackdown against Muslim minorities in Xinjiang), its site says. Great Wall has also worked on a border control and surveillance system that involves the use of Oracle databases, according to Chinese state news agency Xinhua.
Another Chinese reseller, Sinobest, advertises on Oracle’s Cloud Marketplace a “police integrated information system” that runs on the U.S. corporation’s technology. Sinobest describes the app as suitable for “strike, prevent, manage, control” public security, a form of heavy-handed and data-driven policing. “That means that the technology is used to cast a very wide net that covers not only the criminal activity but also every other behavior or condition that is conducive to the crime,” said Daniel Sprick, a legal scholar at the University of Cologne who studies policing in China.
“The technology is used to cast a very wide net.”
Still another partner, Kingbase, is described on a featured page on Oracle’s site as having products that are used in the “national defense war industry.” And a fifth, Xinjiang Hui Wen Network Information Technology, works extensively with the Xinjiang Production and Construction Corps, the entity that runs many internment camps used in the systematic persecution of Muslim Uyghurs and other minorities. Oracle previously told The Intercept that it sold technology to the Xinjiang Public Security Bureau, which oversees policing in the region, until 2019. That was the year the agency was listed by the U.S. Commerce Department.
Moore told The Intercept that Kingbase and Sinobest are not current Oracle partners, despite their pages on Oracle’s website. She said that Oracle does work with Great Wall, the CEC subsidiary, but that the relationship is permitted under the terms of the Defense Department listing. She added that Oracle has no record of police or military transactions involving Great Wall but that the broker has represented Oracle in business with China’s State Administration of Taxation.
Moore said Xinjiang Hui Wen was a partner until earlier this month, when its membership expired, and Oracle is now reviewing the broker’s partner eligibility.
Digital China did not respond to emailed questions, and no one answered when The Intercept called on multiple occasions. Great Wall, Kingbase, and Sinobest did not respond to questions sent through email and submitted through Sinobest’s website. A person who answered the phone at Xinjiang Hui Wen claimed that the reseller did not have an employee who answers media queries. The broker described Oracle as a partner in a June 2020 job listing.
Oracle is not alone among U.S. tech companies in its reliance on resellers in China. For decades, the tech giants have used brokers to provide needed connections to Chinese businesses and government officials.
Oracle’s work with brokers illustrates the role that Western companies play in driving surveillance in China.
Over the past eight years, third-party outfits have become even more important. In 2013, revelations from National Security Agency whistleblower Edward Snowden that U.S. tech companies assisted the agency’s spying set off alarm bells within the Chinese government, sparking a scramble to purge Chinese government supply chains of foreign technology — or at least technology under direct foreign control. To cling to market access, Western companies moved closer to brokers. Oracle, for its part, allowed a Digital China subsidiary to “localize” its powerful server hardware, which is tailored to Oracle databases, resulting in a machine manufactured by the subsidiary. The server was recently sold to Beijing police as part of a cloud-based surveillance project.
Oracle’s work with brokers illustrates the role that Western companies play in driving surveillance in China, even as they scale back their presence there. “The West has a long history of selling arms and surveillance systems to other countries,” said Maya Wang, a senior researcher on China at Human Rights Watch. “There is this pretense that this is a story of good versus evil, with the U.S. being on the good side. But the evidence is not supportive of that narrative.”
Photo: Imaginechina via AP
Brokers like Digital China often work with multiple foreign companies. Digital China’s partners also include Amazon Web Services, IBM, and Microsoft. (In a preliminary search, The Intercept could not find policing-related solutions listed by Digital China on other foreign companies’ websites.)
Ken Glueck, an Oracle vice president, argued that Oracle’s practices in China and elsewhere adhere to the law. “We go beyond what one might anticipate from export control regulations,” he said in a telephone interview with The Intercept. “We vet partners, and we have a track record globally of ending partner relationships where there has been some violation in our view.”
“We have no doubt that all Chinese partners conduct objectionable business.”
Moore, meanwhile, downplayed Oracle’s ability to influence partners. “We have no doubt that all Chinese partners conduct objectionable business,” she said in March, adding that Oracle can only control transactions involving its work. In another email she wrote: “These are arms-length relationships that are almost always far more limited in practice, and by contract, than they may be represented on a public marketing website.”
But at Oracle China’s anniversary event with Digital China, Li, the managing director, reportedly described not an arm’s-length relationship but “mov[ing] forward hand in hand” — a comment underscored by the overlapping hands on the cake knife.
The Rise of Resellers
Foreign technology companies operating in China have long faced a dilemma: In order to obtain market access, they have to give something up. In some cases, that something is their technology, which leaks out through the establishment of joint ventures with Chinese companies and local research and development centers. In other cases, it is their stated values. And often, it is both.
Oracle entered China early, in 1986, and by 2008 it had four R&D centers in the country, according to a presentation found on its website. Ellison, the Oracle co-founder, reportedly selected the location in Beijing himself. The company established a fifth center in 2013 and closed its Beijing center in 2019. Oracle would not say how many of the other centers remain open.
By the early 2000s, Oracle had made significant inroads in the country. The Chinese government was pushing for local control over key technologies, but at the time there was no quality Chinese-made alternative to Oracle’s pioneering database software. Meanwhile, local officials building up data infrastructure were dazzled by brand names, and Oracle was the gold standard in the relational databases used to process large amounts of information. As a result, the Chinese government put Oracle’s technology to use even in sensitive areas. According to a presentation on Oracle’s website, its technology helped power the internet surveillance project Golden Shield, along with the pilot for China’s grid management system, a network of neighborhood-level social control. Golden Shield “was instrumental in helping Public Security Bureaus to identify Uyghurs without proper registration and force them to return to Xinjiang, where many were detained,” said Darren Byler, an anthropologist at Simon Fraser University who studies policing in Xinjiang.
A later Oracle document claims that Oracle servers have been used by the People’s Armed Police, a Chinese paramilitary force focused on domestic security. When asked about Oracle documents claiming business with military entities, Moore said, “We deny transactions for any unlawful or unauthorized military work, consistent with the export laws and regulations at the time of the original transaction, and applicable to any ongoing provision of product support.”
In 2012, company President Mark Hurd told investors on an earnings call that Oracle was “ramping up” its business in Asia. Then the Snowden documents threw Oracle’s China work into limbo. From a hotel in Hong Kong, Snowden helped reveal that the NSA had obtained access to the systems of Apple, Facebook, Google, and other companies. The news spooked Chinese government officials, who worried that a reliance on U.S. technology made China vulnerable to American spying.
“There was increased pressure on foreign companies like Oracle.”
Although it was not named in the documents that Snowden leaked on PRISM, the NSA project, Oracle came under scrutiny because its databases were so widely used by government ministries. “Sometimes in the United States we forget how much the Snowden incident influenced Beijing’s thinking about cybersecurity and policy,” said Elsa Kania, an adjunct senior fellow at the Center for a New American Security and an expert on Chinese military strategy. “In the wake of ‘Prism-gate,’ as PRC state media calls it, there was increased pressure on and suspicion of foreign companies like Oracle.”
Ellison did not help the situation. At a moment when other U.S. tech companies were trying to distance themselves from the NSA, he defended the agency to then-CBS News host Charlie Rose. “It’s great,” Ellison said in the August 2013 interview on NSA surveillance. “It’s essential.” He also asked, “Who’s ever heard of this information being misused by the government?” even though Americans were reeling from a decade of overzealous post-9/11 investigations.
A few days after the interview aired, China’s Ministry of Public Security moved to investigate Oracle, along with IBM and computing company EMC. What ensued was a campaign to “de-IOE” Chinese supply chains, which meant removing IBM, Oracle, and EMC technology — or at least laundering them through Chinese resellers.
As U.S. companies struggled to hold on to market access in China, they made further compromises. Microsoft opened a “transparency center” in Beijing that allowed source code review for its products. IBM partnered up with the Beijing-based broker Teamsun to develop a local database, handing over access to the source code for its Informix software as part of the deal. And Oracle partnered with Digital China to adapt its hardware to the Chinese market. Working with another Chinese company, Neusoft, Oracle also tried to break into the “smart city” business in China, a move reflected in documents found on its site that detail data-driven surveillance as part of a smart city model.
For companies counting on China for growth, the “de-IOE” campaign was far from ideal. But the resellers empowered by it offered one advantage: They helped shroud sales of controversial technologies, warding off scrutiny from both the U.S. government and human rights activists concerned about the ways that Western technology was being used to perpetrate techno-authoritarianism in China. “We’ve seen brokers crop up and become more important as a way to keep things underground,” said Emily Weinstein, a research analyst at Georgetown’s Center for Security and Emerging Technology who has written about problematic work by U.S. tech companies in China.
As U.S. companies struggled to hold on to market access in China, they made further compromises.
Brokered transactions are not totally underground, however. The Chinese government publishes detailed procurement notices outlining the technology that various ministries and departments purchase. Along with police reports, these say that police in Guangdong,Jiangsu, and Yunnan provinces have recently purchased Oracle databases. So has China’s Ministry of Public Security, which oversees police throughout the country. An online marketplace set up by China’s State Council notes, “Oracle databases are used in various fields in China, such as e-government systems at all levels of government.” Among other areas, it cites public security.
Procurement records show that Digital China or its offshoots have licensed Oracle databases, applications, and middleware to several government clients, including the State Administration of Radio, Film, and Television, the former name of one of the entities responsible for censorship in China. Some of these deals have been lucrative. A 2018 project to provide Oracle database software to the central government’s State Sports General Administration, for example, brought in close to 19.8 million RMB ($2.9 million) for Digital China. It is unknown what portion of that went to Oracle.
Digital China also brought in money for Oracle through another source: the localized Oracle database server, or Dengyun All-in-One.
Oracle Servers, Made in China
Specifications given by a Digital China subsidiary for the Dengyun server (right/bottom) are identical to those listed by Oracle for its Exadata X8-2 server (left/top).Images: Oracle
The origin of the Dengyun All-in-One is detailed in a presentation hosted on Oracle’s website, which says that the tech giant and Digital China signed an agreement covering the transfer of Oracle hardware in September 2016. Under the deal, Oracle agreed to assist a Digital China subsidiary, Yunke China, with rolling out the server, a towering machine designed to run Oracle’s relational databases. Oracle was to provide marketing, project management, and some technical support, according to an account published on Digital China’s site. Oracle marketing documents tout the server as suitable for handling sensitive data and cloud-ready. (The “yun” in Dengyun means “cloud.”)
Moore downplayed the extent of the collaboration. “The Digital China Dengyun offering is simply a rebranding of older versions of Oracle’s hardware,” she said. “It involves no joint development or intellectual property transfer of any kind.” Oracle said that Dengyun was made of “parts” from Oracle’s Exadata Database Machine, a server that it markets to businesses and government clients in the United States and elsewhere as “the best platform to run the Oracle Database.” The specifications given by Yunke China for the most recent Dengyun All-in-One are identical to those that Oracle lists for the Exadata X8-2.
Within four months of Oracle signing the Dengyun agreement, assembly lines in a factory in Guangdong province were churning out the machines, according to the Oracle presentation and to Yunke China’s site. That same month, provincial government social security departments in Hebei and Shaanxi provinces received the first Dengyun servers, the presentation says.
The following year, Yunke China displayed one of the servers at an Oracle conference, according to the broker’s website.
It is unclear exactly how Dengyun All-in-One intersects with Oracle’s efforts to market its data analytics software for use by Chinese police, but the product helped position Oracle and its partners to upsell customers on advanced software that works with its database. A source close to the broker industry said that companies like Oracle frequently pitch additional products to existing database clients.
What is clear is that for Oracle, the Dengyun deal meant its databases would remain ubiquitous and necessary at a moment when its China business was in danger. The tech giant appears to have continued to assist Digital China and other brokers with sales. Dengyun is mentioned in nearly a dozen Oracle marketing documents downloaded from the tech giant’s website.
A timeline in an Oracle document shows the company’s collaboration with Digital China, ending with a database machine that it boasts is made in China.
Dengyun’s local connection became part of the sell. A 2016 Oracle document depicts the company’s collaboration with Digital China on a timeline, mentioning the “domestically manufactured” Dengyun at the end of the trajectory. A later sales document pitching Oracle’s technology to Chinese government cloud clients touts Dengyun as “autonomous and controllable,” a term widely used in China to mean Chinese-made products that have no direct oversight by foreign entities and no backdoors. (The Intercept downloaded the second presentation before it disappeared from Oracle’s website.)
By the time of the Dengyun deal, Digital China was enmeshed in surveillance work. In a post dated August 2018, shortly before it and Oracle celebrated 20 years of teamwork, Digital China announced on its website a geospatial effort called Deeplan. The project uses mobile phone data, apparently culled from telecommunications giant China Mobile, to generate heat maps showing potential crowds. The goal is to “monitor the activities of people in real time” in order to “improve the efficiency of police officers,” according to the site. An apparent image from the interface shows individuals’ blurred-out names, along with their phone numbers, precise locations, and cities of origin. The page says police can use the heat map to divide people into groups from inside and outside a province, a feature that might be used to track aspiring protesters or government petitioners. The Digital China subsidiary that oversees the project, Howso Technology, currently lists a job opening on its website for a software development engineer specializing in geographic information systems, or GIS. Among the requirements is familiarity with Oracle databases.
Do you have information on how Oracle products are used in China or in security applications? Have you worked with the company? Contact the author of this story, Mara Hvistendahl, on Signal at +1 651-400-7987 or at email@example.com.
The Dengyun effort was soon linked to oppressive entities in China as well. In 2018, the server was sold to the National Geomatics Center of China, which oversees geospatial work that is crucial for police efforts.
Most glaringly, on March 9 the Beijing city government announced that it would pay 2.7 million RMB ($414,900) for a Dengyun machine as part of a “smart policing project.” Although ostensibly focused on transportation, the project is part of China’s “Police Cloud,” an effort to collect and centralize information that has been criticized by Human Rights Watch. (The Intercept previously reported that Oracle employees promoted the company’s technology for Police Cloud.)
A procurement document for the project describes features of interest to police trying to prevent protests or large gatherings, including automated license plate recognition, the identification of cars that are entering the city for the first time, and facial recognition on drivers inside their vehicles. It also describes the server as a necessary part of the policing project and notes that the project requires the purchase of certified Oracle databases, as well as “at least one Oracle-certified engineer.”
Glueck, the Oracle vice president, claimed that the Dengyun server is too basic to be of use in any kind of ambitious surveillance project. “This product is not the kind of appliance that you could even run a surveillance system on,” he said. “That’s not to say that we obviously don’t have different hardware capabilities, but this is a very limited, small-scale kind of appliance.”
But Oracle’s own marketing materials boast that Dengyun can be used for big data and high-performance computing. The Yunke China flyer separately says that the server offers 3.0 petabytes of disk capacity per rack — enough to store years’ worth of continuous high-definition video or billions of social media images.
Oracle’s relationships with resellers are a cautionary tale. In its effort to distance itself from the U.S. government in China, the company may have gone too far in the opposite direction.
Enabling Data-Driven Policing
For many Western companies operating in China, brokers offer a sort of plausible deniability, allowing them to claim that they simply provide neutral tools and don’t bear responsibility when those tools are used to surveil the Chinese population. But it is hard for Oracle to make that claim with Digital China: The broker markets policing-oriented products configured for Oracle technology on the tech giant’s own site.
On its partner page on Oracle’s site, Digital China says it can help Chinese authorities maintain a data-driven policing system.
One such offering on Digital China’s Oracle.com partner page is called “Government Industry Information Security Solution 2.0.”
A Chinese-language posting about the “solution” says that Digital China can provide IT maintenance and management for Chinese police, military, and other entities that conduct surveillance. This includes work for China’s 1203 Project, an effort to link up security data infrastructure from across China. The partner page also describes work for the feared chengguan, uniformed city management officers who are known for their violence toward street vendors. And it boasts that Digital China can help authorities maintain an “integrated policing platform,” a term used to describe data-driven policing software. The page notes that the solution is built to run on Oracle databases, operating systems, and virtual machines.
“It clearly shows that Oracle did not shy away from providing crucial technological infrastructure to the Chinese party-state,” said Sprick, the China legal scholar, after reviewing the page.
“We all have different definitions of what surveillance might be.”
Digital China is an Oracle managed service provider, a coveted certification that the tech giant says requires annual third-party audits. In a February email, Moore said that the company separately audits or investigates any partners suspected of misusing its technologies. Oracle declined to comment on what its audits in China entail. Glueck claimed that Oracle has ended partnerships in the country based on audit results but would not give details.
Byler, the anthropologist, said companies like Oracle need to do better. “It does take some time to suss out your networks and figure out who is doing what,” he said. “But due diligence and the human rights record of China demands that you do that kind of work.”
When asked about the police assistance described on Digital China’s partner page, Moore equivocated. Rather than disavow the project, she said, essentially, that it wasn’t that bad. She noted that the solution includes other, nonpolicing government work. “An examination of the full text shows the offering you cite is almost entirely ‘optimizing’ other solutions, nearly all of which are completely benign,” she said, adding, “If you object to ‘video conference optimization,’ we really don’t know how to respond.” She neglected to note that the solution describes video conference optimization work for Chinese military clients — for an unspecified “army office” in the “national defense industry.”
“We all have different definitions of what surveillance might be,” Glueck told The Intercept. But when pressed for Oracle’s definition, he diverted the conversation. Instead of answering, he suggested looking into the tech giant’s rivals.