A look at the many ways China suppresses online discourse about the Tiananmen Square protests

Comment

People hold candles during a vigil in Hong Kong on June 4, 2018, to mark the 29th anniversary of the 1989 Tiananmen crackdown in Beijing. - Crowds assembled in Hong Kong on June 5 in memory of the victims of China's Tiananmen Square crackdown but young activists are increasingly questioning the annual vigil's relevance. Semi-autonomous Hong Kong has seen tens of thousands gather at the candlelit vigil in Victoria Park since 1990, while any mention of Beijing's brutal crackdown on students calling for democracy on June 4, 1989, remains strictly censored in the mainland. (Photo by Anthony WALLACE / AFP) (Photo credit should read ANTHONY WALLACE/AFP/Getty Images)
Image Credits: ANTHONY WALLACE/AFP/Getty Images (opens in a new window) / Getty Images

Every year before the anniversary of the Tiananmen Square Massacre on June 4, the Chinese government begins to exert even more control over what information people can access online. This year is the 30th anniversary of the day the protests ended when the government sent troops to fire on students, activists and other people demonstrating against the Communist Party. Censorship efforts appear to have intensified, potentially affecting even social media accounts outside of China.

From the beginning, the government has tried to control information about the massacre, claiming after the event that 300 people died, though Amnesty International and other international observers said the death toll may have been much higher, potentially numbering into the thousands.

Suppression of information means that an entire generation of people know little about the events, even as the activists involved continue to suffer repercussions, including long prison sentences. In recent years, the government’s censorship apparatus has become even more powerful, with voice and image recognition and machine learning making it easier to block or remove posts at scale. As a content screening employee at Bytedance recently told Reuters, “we sometimes say the artificial intelligence is a scalpel, and the human is a machete.”

This year, one of the most notable examples of the government’s annual crackdown on information began in April, when every language version of Wikipedia was blocked in China, instead of just the Chinese-language version and individual articles about sensitive issues like the Tiananmen Square protests and Tibet.

In May, video-sharing sites Bilibili and AcFun suspended real-time comments, while Douban E Zu (the name translates to “Douban Goose Group”), a popular celebrity gossip and news forum with more than 600,000 members, halted service between May 30 and June 29. Both services claimed they needed to perform “system maintenance.”

Online discourse is already strictly controlled by the Chinese government, which requires all websites to do real-name checks on users when they register an account (for example, by linking phone numbers, which are tied to government-issued IDs). Discussions on Douban E Zu often center around politics, which may have prompted heavier restrictions. Real-time comments (called “bullet screens”) on Bilibili and AcFun are harder to monitor for banned content and even though the government recently issued new guidelines for screening comments on bullet screens, censors may still be working on ways to maintain control on them.

Most recently, WeChat, the ubiquitous messaging, games and e-commerce platform, blocked users from changing their headshots, alias and What’s Up status. Then this weekend, users began reporting connection issues with their VPN services, which are used to get around mainland China’s “Great Firewall” and access forbidden sites.

The effects of the crackdown on information about Tiananmen Square also appears to have spread beyond China. Yaxue Cao, founder and editor of China Change, a U.S.- based English-language about human rights and law in China, tweeted over the weekend that many Chinese-language Twitter accounts critical of China had been suspended, even if they originated from outside of China, including some based in the U.S.

Twitter claimed on its Public Policy account that it had suspended accounts mostly for “engaging in a mix of spamming, inauthentic behavior, & ban evasion.” Addressing the suspension of accounts that “were involved in commentary about China,” Twitter’s Public Policy account claimed that contrary to speculation, “these accounts were not mass reported by the Chinese authorities—this was a routine action on our part. Sometimes our routine actions catch false positives or we make errors. We apologize. We’re working today to ensure we overturn any errors but that we remain vigilant in enforcing rules for those who violate them.”

Several of the accounts Cao mentioned, including @Sasha_Gong and @wmeng8, have since been restored, but users were frustrated by Twitter’s explanation, especially since the suspensions affected so many Chinese dissident accounts, and called on the company to investigate potential causes.

A look back at years of Tiananmen Square censorship

An earlier example of Chinese government attempts to suppress information about Tiananmen Square each year took place in 2010, when Foursquare was blocked, possibly as a response to users checking in to Tiananmen Square to commemorate the anniversary.

The Chinese government also reduced access to sites that were once at least somewhat accessible is Tumblr, which in 2016 joined the ranks of other social media sites blocked in China. GreatFire.org’s historical data shows that Tumblr became more difficult to access in June 2014 and 2015, but in 2016 that appeared to become permanent, with GreatFire.org’s tests showing the site 100 percent blocked since then (perhaps because of the Chinese government’s ban on online porn).

For foreign companies that want to be allowed to do business in China, that means capitulating to its censorship laws even at the risk of angering their users in other countries. In 2014, LinkedIn, one of the few U.S.-based social media sites still allowed in China, said it would not only prevent users in China from seeing content prohibited by the government, but also block users in other countries from seeing banned content that originated in China. The decision was criticized after members noticed that posts relating to the Tiananmen Square anniversary had been blanked out, reported the Guardian.

More recently, LinkedIn backed down after blocking the account of activist Zhou Fengsuo in China. Zhou, who was one of the student leaders of the Tiananmen Square protests, told the South China Morning Post in January that he believed the censorship had more to do with government monitoring of his activism on other social media platforms, like WeChat, than his LinkedIn content. After restoring Zhou’s LinkedIn profile, the company claimed it had been “blocked in error.”

Another case of Chinese government censorship affecting even users in other country took place in 2017, when microblogging service Weibo banned overseas users from uploading images and videos during the week of the anniversary. As in the case of Bilibili and Douban E Zu this year, Weibo claimed that the action was related to a systems upgrade.

More TechCrunch

ClickHouse has made a name for itself as a real-time data warehouse for large enterprises. Its customer list includes Deutsche Bank, eBay, Fastly, GitLab, HubSpot, Microsoft, ServiceNow and Spotify.

Real-time database startup ClickHouse acquires PeerDB to expand its Postgres support

The EU has kicked off a consultation on rules that will apply to providers of general purpose AI models under the bloc’s AI Act.

EU calls for help with shaping rules for general purpose AIs

Siddhi Capital’s second fund of $135 million is double the size of the venture capital firm’s first fund and will go into CPG and food tech startups.

Siddhi Capital grabs $135M for Fund II to invest in consumer packaged goods startups

Perplexity AI will soon start sharing advertising revenue with news publishers when its chatbot surfaces their content in response to a user query, a move that appears designed to assuage…

Perplexity details plan to share ad revenue with outlets cited by its AI chatbot

Femtech, or tech that leverages innovations in AI, smartphones and connected wearables to give women more insights into reproductive and menstrual health, continues to gain momentum with users, and investors…

Fertility tracking app Flo Health raises $200M at a $1B+ valuation

Meta said Monday that it is rolling out its AI studio to all creators in the U.S. to let them create personalized AI-powered chatbots. The company first announced the AI…

Meta is rolling out its AI Studio in the U.S. for creators to build AI chatbots

The startup is betting it can bring its fusion technology to market at a breakneck pace by leaning heavily on partners.

Bill Gates-backed Type One Energy lands massive seed extension to commercialize fusion power

The U.K.’s antitrust regulator has revealed an early-stage probe into Google’s ties with Anthropic, after the Alphabet subsidiary invested in its U.S. AI rival over several rounds. While it’s not…

UK antitrust body probes Google’s ties with AI rival Anthropic

We’ve been covering US-based Insurtech startup FAYE way back since 2022 with its Seed round, and the Series A round in 2023, and it seems they continue to be on…

Packing travel insurance products into an app helped FAYE to a $31M Series B

Sennder is acquiring the European ground transportation assets of logistics giant C.H. Robinson.

Sennder buys CH Robinson’s European business

When Egyptian B2B e-commerce platform Cartona last raised money in 2022, global and local investors were eager to invest in African startups solving the supply chain and operational challenges for…

Egypt’s Cartona raises $8.1M even as investors pull back from B2B e-commerce in Africa

During an emergency hearing held by the South Korean government, Young-bae Ku, Qoo10’s founder said he would secure the amount over the next 30 days.

Qoo10’s CEO pledges personal assets worth $58M to compensate Korean merchants affected by its liquidity crisis

Zoe sends customers at-home testing materials to collect blood or feces to test blood fat, blood sugar, and gut microbiome health. Following those results, the company scores every food (on…

Zoe, a microbiome-focused nutrition company, raises $15 million to expand in the U.S.

Qualcomm has launched the Snapdragon 4s Gen 2 chip to 5G smartphones in the sub-$100 price segment in India and other emerging markets.

Qualcomm’s new Snapdragon chip aims to bring 5G to sub-$100 devices

Users who have signed up for iOS 18’s developer beta can now get the first taste of Apple Intelligence as the company has released some features of its AI suite…

Apple brings Apple Intelligence with iOS 18.1 dev beta, but there are a lot of limitations

Canva has acquired Leonardo.ai, a generative AI content and research startup, as the company looks to broaden the scope of its AI tech stack. The financial terms of the deal…

Canva acquires Leonardo.ai to boost its generative AI efforts

The U.S. Commerce Department has issued a new report endorsing open models like Meta’s Llama 3.1, which it says promote competition.

New U.S. Commerce Department report endorses ‘open’ AI models

Shared micromobility giant Lime is piloting two new vehicles designed to appeal to women and older folks who might appreciate a lower step-through frame, smaller wheels and an upgrade from…

Lime is piloting two new e-bikes to attract more women and older riders 

Apple has published a technical paper detailing the models that it developed to power Apple Intelligence, the range of generative AI features headed to iOS, macOS and iPadOS over the…

Apple says it took a ‘responsible’ approach to training its Apple Intelligence models

A fireside chat on Monday between Nvidia CEO Jensen Huang and Meta CEO Mark Zuckerberg at the SIGGRAPH 2024 conference in Colorado took a few unexpected turns. It started innocently…

Huang and Zuckerberg swapped jackets at SIGGRAPH 2024 and things got weird

Meta’s machine learning model, Segment Anything, has a sequel: It now takes the model to the video domain, showing how fast the field is moving.

Zuckerberg touts Meta’s latest video vision AI with Nvidia CEO Jensen Huang

Featured Article

The fall of EV startup Fisker: A comprehensive timeline

Here is a timeline of the events that led fledgling automaker Fisker to file for bankruptcy.

The fall of EV startup Fisker: A comprehensive timeline

Hello, and welcome back to TechCrunch Space. In case you missed it, Boeing and NASA decided to keep Starliner docked to the International Space Station for the rest of the…

TechCrunch Space: Catching stars

As failed EV startup Fisker winds its way through bankruptcy, a persistent and tricky question has become a flashpoint of the proceedings: does its only secured lender, Heights Capital Management,…

The question haunting Fisker’s bankruptcy

So-called “unlearning” techniques are used to make a generative AI model forget specific and undesirable info it picked up from training data, like sensitive private data or copyrighted material. But…

Making AI models ‘forget’ undesirable data hurts their performance

Uber is now letting riders in India book up to three rides simultaneously.

Uber now lets users in India book three trips at once

U.S. airports are rolling out facial recognition to scan travelers’ faces before boarding their flights. Americans, at least, can opt out. 

How to opt out of facial recognition at airports (if you’re American)

The promise of AI and large language models (LLMs) is the ability to understand increasingly wider amounts of context and make sense of that information easily, so it makes sense…

Bee AI raises $7M for its wearable AI assistant that learns from your conversations

Featured Article

DEI backlash: Stay up-to-date on the latest legal and corporate challenges

It’s clear that this year will be a turning point for DEI.

DEI backlash: Stay up-to-date on the latest legal and corporate challenges

Bike-taxi startup Rapido, which counts Swiggy among its investors, is the latest Indian firm to become a unicorn.

India’s Rapido becomes a unicorn with fresh $120M funding