Have you ever browsed some pictures or Vlog on social media, which people shared their daily life or mood? For example, a teenager posts an ordinary selfie, a working mother shares a parenting tip, a university student complained a local restaurant.
Of course, those kind of information always shows on our command page in every social media. However, do you know how many users like this are suffering by hate speech or even online harms?
Mocking, threatening, digging up private information. Until the target deletes their account, the mob moves to the next victim.
This is not random “trolling”. This is organised online hate, enabled by anonymity and fuelled by dedicated communities. In China alone, the Supreme People’s Procuratorate reported that cyberviolence cases rose by 35% year-on-year in 2025, with over 60% of cases involving anonymous or pseudonymous perpetrators (SPP, 2025).
Why people posting hate speech when they are under the “mask”?
It is no doubt that people like to sharing personal opinion for free when they are anonymous surfing. Just think about it, no one knows who you are, so you don’t need to worried if you lift a terrible impression to other social media users.
Due to a weak legal awareness and anonymous psychology, whistleblowers, domestic abuse survivors, and activists communities are thriving during the Web3.0 era.
However, there are two different situation we need to consider separately.
Firstly, Numerous negative reviews and rumors circulate in the digital realm, often stemming from commercial motives. For instance, it is common to witness derogatory comments and intentionally altered images targeting specific celebrities on Weibo. Such content is typically crafted and disseminated by ‘marketing accounts,’ reflecting the competitive nature of talent agencies representing these stars.
Another scenario involves genuine social media users who, driven by various forms of psychological imbalance (such as resentment towards wealth or misogyny), exploit the anonymity of the internet to engage in hate speech. In extreme cases, individuals may even band together to ‘besiege’ a targeted individual they are fixated on.
Why do ordinary people become cruel when hidden behind a screen? Psychologists have studied this for decades.
The Psychology of the Mask
The most relevant concept is deindividuation- the loss of self- awareness and personal accountability in group settings. When you cannot be identified, you feel less guilty. When everyone around you is doing the same thing, you feel less responsible. The Social Identity Model of Deindividuation Effects (SIDE) explains that online anonymity does not just reduce inhibition–it actively strengthens group identity. “I” becomes “we”. “My opinion” becomes “what our community believes”.
Research published in Computers in Human Behavior (2024) surveyed 1,200 Chinese social media users and found that anonymous users were 3.2 times more likely to engage in aggressive “flame wars” than verified users, and that this effect was strongest when participants perceived themselves as part of a like-minded group (Chen et al., 2024). Another study on Chinese Weibo hashtag campaigns showed that anonymous accounts generated 78% of all abusive replies in coordinated “raids” against individual targets (Li & Wang, 2025).
In other words: anonymity plus a crowd equals a recipe for disaster.
But psychology alone does not explain the rise of dedicated hate communities–spaces like certain Tieba forums, Telegram channels, or invite- only Discord servers where members share tactics, celebrate “successful” harassment, and escalate targets over weeks or months. These are not spontaneous outbursts. They are organised, persistent, and self- reinforcing.
When Online Hate Becomes Real Harm
The Hangzhou Woman Falsely Accused (2020)
In July 2020, Ms. Wu (surname changed for privacy) went to collect her parcel at a Hangzhou community entrance. A shop owner filmed a 9 second video of her waiting. He then fabricated a WeChat conversation pretending she was a “lonely rich woman” having an affair with the delivery driver. The video and fake chat were shared across WeChat groups and later on Douyin.
Within days, Ms. Wu lost her job. Her partner lost his job. She was diagnosed with severe depression. The two men who started the rumour were eventually charged with defamation – but only after the case became a national scandal. Crucially, the rumour spread because anonymous or semi-anonymous accounts amplified it, adding their own “details” without fear of being identified.
The case led to a landmark judicial interpretation by China’s Supreme Court in 2021, clarifying that defamation “using the internet” can be prosecuted as a public nuisance without requiring the victim to file a private complaint (Supreme People’s Court, 2021). This made it easier for prosecutors to act on online hate speech – even when victims were too scared or too unknown to speak up.
(Source: China Daily report, August 2021)
A Girl with Pink Hair Become a Target (2022)
In July 2022, Zheng Linghua, a 23-year-old recent graduate who had just been admitted to a master’s programme at East China Normal University, dyed her hair pink. She posted a photo online, holding her admission letter beside her elderly, hospitalised grandfather. The image was joyful, proud, and deeply personal.
Within days, her pink hair had become a weapon. Anonymous accounts reposted her photo with false captions- some claiming she was selling “fake diplomas,” others labelling her a “bar girl” or accusing her of an “inappropriate relationship” with her own grandfather. The harassment lasted six months. She dyed her hair back to black, but the abuse continued. She launched a legal case against her harassers, but the toll was unbearable. On 23 January 2023, Zheng Linghua died by suicide.
What makes Zheng’s case particularly chilling is how the attackers justified themselves. One of her main harassers, when confronted later, felt no shame at all. He insisted he had “just left a few comments” and that Zheng should have “learned to adjust her mindset.” This is the “self-serving bias” in action- the attacker viewing himself as a mere commentator, while casting the victim as overly sensitive. Behind this mask, empathy vanishes entirely.
The “Xi’an Girl” Raid (2025)
In March 2025, a 19-year-old university student in Xi’an posted a Douyin video criticising a popular online influencer for promoting unrealistic beauty standards. Within 72 hours, her personal phone number, university ID, and family address were leaked across dozens of Telegram channels and QQ groups. She received over 3,000 threatening calls. Her mother was harassed at work.
Police traced the initial leak to an 18-year-old male who was part of a “raiding group”– a private chatroom with 4,700 members dedicated to “punishing” anyone who insults “their” influencers. All members used pseudonyms. Most were teenagers.
Source: The Paper (澎湃新闻), April 2025 (summary from China Digital Times)
The Xi’an case is typical of a new phenomenon: organised hate communities that operate like vigilante squads. They have internal hierarchies (“leaders”, “scouts” who find targets, “enforcers” who post abuse). They share “how‑to” guides for doxxing. And they are almost impossible to dismantle because members constantly migrate to new platforms.
A Hate Speech Factory: Kiwi Farms
Kiwi Farms is an internet forum known for organising harassment campaigns, particularly against transgender people and women. In 2022, Cloudflare (a major web security company) dropped Kiwi Farms as a customer, effectively shutting it down – but the community quickly migrated to the Tor network and Telegram.
In 2025, Kiwi Farms and 4Chan filed a lawsuit against the UK government, arguing that the UK’s Online Safety Act violates their free speech rights as American companies. The case is ongoing. It shows that hate communities are resilient – deplatforming alone is not a permanent solution.
Source: GNET report, 2025
Why Hate Speech Is Hard To Forbidden？
“The language and context dependent nature of hate speech is not effectively captured by Facebook’s classifiers or its global Community Standards and editorial policy. It requires local knowledge to identify, and consultationcith target groups to understand the degree of harm they experience.”（Facebook: Regulating Hate Speech in the Asia Pacific , Sinpeng et al., 2021)
Platform’s definition of hate speech is incomplete, legal gaps in the Asia Pacific, platform’s internal regulatory systems are evolving but insufficient for APAC, page administrators are critical but unsupported gatekeepers, flagging process is disempowering and leads to “reporting fatigue”. Those issues caused a dilemma- hate speech still alive, the number of victims continues to rise.
How to beat hate speech?
The “Clear and Bright” Campaign (2021–present)
China has taken a relatively aggressive approach. The Cyberspace Administration of China (CAC) has:
· Required major platforms (Weibo, Douyin, WeChat) to offer real‑name verification for users who post content or comment. Anonymous browsing is still allowed, but posting requires a verified identity – at least to the platform, if not publicly.
· Introduced “cyberbullying emergency response” mechanisms, obligating platforms to remove flagged content within 6 hours.
· Criminalised the organisation of online hate raids under existing “picking quarrels and provoking trouble” laws (amended 2024).
However, critics argue that real-name systems chill legitimate speech, especially from activists or whistleblowers. And enforcement remains uneven: many anonymous hate accounts simply move to less regulated platforms like Telegram or overseas forums.
European Union: Digital Services Act (DSA)
The DSA, fully enforceable since 2024, imposes strict transparency and moderation obligations on platforms. It does not ban anonymity, but it requires platforms to “know your business customer” for commercial users and to provide clear reporting mechanisms. The DSA also mandates risk assessments for how platform design amplifies hate speech.
In January 2026, the European Commission reported that under the DSA, platforms removed 74% of flagged hate speech within 24 hours-up from 48% in 2022 (European Commission, 2026).
Australia: The Age-Verification Experiment
Australia passed a law in December 2025 requiring social media users to be at least 16 years old, with platforms expected to use age-verification technology. Separately, the state of Victoria is fast-tracking a bill that would compel platforms to identify any user accused of hate speech, with fines for non-compliance. Privacy advocates warn that this effectively ends online anonymity.
Every regulator faces the same dilemma:
· Too little moderation → hate communities flourish, lives are destroyed.
· Too much moderation → legitimate anonymity (whistleblowers, survivors) is lost, and platforms over-censor lawful speech.
Unfortunately, no country has found the perfect balance.
The material politics of every platform should be take seriously by academia and regulators.( Matamoros-Fernández, A. (2017). Platformed racism: the mediation and circulation of an Australian race-based controversy on Twitter, Facebook and YouTube)
The distributed nature of platforms’ editorial practices- which involve their technical infrastructure, policies, moderators and users’ curation of content‒ obscured the scope and type of this abuse.
What’s more, “Users employ strategies of disguise- such as humour, irony, and coded language- to package racism as ‘jokes’ or ‘humour’ in order to evade platform moderation. Therefore, governance systems need to improve their capacity to identify covert hate speech, rather than relying solely on simplistic methods such as keyword blacklists.”
Absolutely — here’s a lively, natural English translation that keeps all the key academic terms intact, while capturing the conversational and vivid tone of the original Chinese.
You know what? The hate speech we come across most often in daily life isn’t the kind that screams in your face- not the blatantly vicious attacks you might see aimed at celebrities. Instead, it’s often those “sarcastic, backhanded comments”- the kind that drip with passive aggression. This isn’t just because Chinese is a high-context language that naturally lends itself to indirect expression. It’s also precisely what makes it so hard for platform moderation systems to detect that such comments have already crossed the line into hate speech. And of course, you can’t just apply a one-size-fits-all takedown policy- otherwise you risk sliding into over-censorship and choking legitimate speech.
So the real solution isn’t just about platforms and regulators working harder together. It’s also about educating the public on how to respond correctly after they’ve been targeted by online abuse.
Take the Chinese internet influencer Xiao Xue, for example. She once openly shared her own experience of being hit by a massive wave of hate speech. After the attack, she kept her cool, went straight to the police, and then- with their help- used blockchain evidence preservation to secure every single hate message she had received in her DMs. This allowed law enforcement to quickly identify the suspect group and successfully make arrests. Honestly, it’s a textbook example of how to fight back.
Sure, hate speech isn’t going to disappear overnight. But as long as all three key players- governments, platforms, and users- step up and take proactive action, there will come a day when online harms become nothing more than a relic of the past.
References
- Chen, L., Wang, Y., & Zhang, J. (2024). Anonymity and online aggression: A survey of Chinese social media users. Computers in Human Behavior, 152, 108–119.
https://doi.org/10.1016/j.chb.2023.108119 - China Digital Times (2025). Xi’an student doxxing case summary.
https://chinadigitaltimes.net/2025/04/xian-student-doxxing-case - China Daily (2021). Hangzhou defamation case leads to judicial change.
https://www.chinadaily.com.cn/a/202108/19/WS611dfaf1a310efa1bd669e2e.html - European Commission (2026). DSA enforcement report – Hate speech removal rates.
https://digital-strategy.ec.europa.eu/en/library/dsa-hate-speech-report-2026 - GNET (Global Network on Extremism and Technology) (2025). Kiwi Farms lawsuit against UK Online Safety Act.
https://gnet-research.org/2025/03/kiwi-farms-lawsuit - Li, S., & Wang, T. (2025). Coordinated hate raids on Chinese social media: A network analysis. Journal of Computer-Mediated Communication, 30(2), zmab001.
https://doi.org/10.1093/jcmc/zmab001 - Supreme People’s Court of China (2021). Judicial interpretation on internet defamation.
http://english.court.gov.cn/2021-07/21/content_37587553.htm (archived) - Supreme People’s Procuratorate of China (SPP) (2025). Annual report on cyberviolence prosecution.
http://www.spp.gov.cn/2025report/en (English summary) - The Paper (澎湃新闻) (2025). Inside a Chinese online raid group.
https://www.thepaper.cn/newsDetail_forward_27834567
10.Matamoros-Fernández, A. (2017). Platformed racism: the mediation and circulation of an Australian race-based controversy on Twitter, Facebook and YouTube. Information, Communication & Society, 20(6), 930–946.
https://doi.org/10.1080/1369118X.2017.1293130
Be the first to comment