Study on Hate Speech on Social Media
Study on Hate Speech on Social Media
社交媒體仇恨言論研究
Introduction
A group of experts looked at how social media sites stop hate speech against Jewish people.
一群專家研究了社交媒體網站如何阻止針對猶太人的仇恨言論。
Main Body
YouTube did not remove a video about a bad attack at Bondi Beach. The video said the attack was a lie. The experts say YouTube's rules are not good enough.
YouTube 沒有刪除一段關於邦迪海灘惡性襲擊的影片。該影片稱這次襲擊是謊言。專家表示 YouTube 的規範不夠完善。
TikTok uses computers to find bad videos. They say they remove 98% of bad content quickly. But sometimes, human workers put the bad videos back online.
TikTok 使用電腦篩選不良影片。他們表示能快速刪除 98% 的不良內容。但有時人工審核人員會將不良影片重新上傳至網路。
One group studied many sites. TikTok removes more hate speech than Meta and X.
有一個小組研究了許多網站。TikTok 刪除的仇恨言論比 Meta 和 X 更多。
Conclusion
The experts will continue to ask questions to stop hate online.
專家將繼續提出質詢，以阻止網路仇恨。
Vocabulary Learning
⚡ Action Words: Now vs. Always
Look at how the story talks about what happens on the internet. We use different 'shapes' for words depending on if it is a general fact or a specific action.
1. General Facts (The 'S' Rule) When we talk about one company, we add an -s to the action:
- TikTok uses computers.
- TikTok removes more hate speech.
2. Specific Past Actions (The 'ED' Rule) When it already happened, we often add -ed:
- Experts looked at the sites.
- YouTube did not remove (Wait! After 'did not', the word stays simple).
Quick Map for A2 Beginners:
Company Name Action + s (Fact)
Past Time Action + ed (History)
Key Words to Steal:
- Remove To take away.
- Continue To keep doing something.
Vocabulary Learning
Royal Commission Investigates How Social Media Platforms Manage Antisemitic Content
皇家委員會調查社交媒體平台如何管理反猶太內容
Introduction
The Royal Commission into Antisemitism and Social Cohesion has held hearings to examine how effectively major digital platforms moderate content to stop antisemitic speech and misinformation.
反猶太主義與社會凝聚力皇家委員會已舉行聆訊，以審查各大數位平台在阻止反猶太言論與錯誤資訊方面的內容審核成效。
Main Body
The hearings revealed a major difference in how different platforms operate. For example, YouTube was criticized for keeping a video that claimed a survivor of the Bondi Beach attack was a 'crisis actor' in a fake operation. Despite claims that the victim was an intelligence agent using makeup, Rachel Lord, Google's head of government affairs in Australia, emphasized that senior reviewers decided the video did not break any rules. Consequently, Richard Lancaster, the Counsel Assisting, described the platform's guidelines as having a 'serious deficiency.'
聆訊揭露了不同平台在運作方式上的重大差異。例如，YouTube 因保留一段聲稱邦迪海灘襲擊倖存者是假操作中「危機演員」的影片而受到批評。儘管有人指稱該受害者是使用化妝的情報員，但 Google 澳洲政府事務主管 Rachel Lord 強調，高級審核員判定該影片並未違反任何規則。因此，協助法律顧問 Richard Lancaster 將該平台的指南描述為有「嚴重缺陷」。
In contrast, TikTok presented a more proactive approach. Zachary Hecht, the global head of policy, testified that the platform uses automated tools to check all uploads, claiming that 98% of violating content in Australia was removed automatically in the first quarter of the year. Furthermore, TikTok activated its 'crisis response protocol' within one hour of the Bondi Beach incident. However, the commission found a problem where human moderators sometimes accidentally restored hate speech that the automated system had correctly removed, meaning external groups had to report the content again to get it deleted.
相比之下，TikTok 採取了更為主動的做法。全球政策主管 Zachary Hecht 證稱，該平台使用自動化工具檢查所有上傳內容，並聲稱今年第一季在澳洲 98% 的違規內容為自動刪除。此外，TikTok 在邦迪海灘事件發生後一小時內便啟動了「危機應對協議」。然而，委員會發現了一個問題，即人工審核員有時會誤將自動系統已正確刪除的仇恨言論恢復，這意味著外部團體必須再次舉報該內容才能將其刪除。
Data from the Online Hate Prevention Institute shows that platforms vary in their responsiveness. According to their analysis, TikTok removed reported harmful content at a higher rate (64%) than Meta (54%) and X (24%).
來自網路仇恨防治協會 (Online Hate Prevention Institute) 的數據顯示，各平台的反應程度各異。根據其分析，TikTok 刪除舉報有害內容的比率 (64%) 高於 Meta (54%) 與 X (24%)。
Conclusion
The commission is continuing its investigation into how digital platforms allow online hate to spread, and more witnesses are expected to testify.
委員會將繼續調查數位平台如何導致網路仇恨傳播，預計將有更多證人出庭作證。
Vocabulary Learning
⚡ The 'Logical Glue': Moving from Simple to Complex Sentences
At the A2 level, you likely use and, but, and because. To hit B2, you need Connectors of Contrast and Result. These words act like glue, making your writing sound professional and academic rather than like a list of simple facts.
🔍 The Shift: From A2 to B2
| A2 (Simple) | B2 (Advanced Bridge) | Why it's better |
|---|---|---|
| But TikTok was faster. | In contrast, TikTok presented a more proactive approach. | It signals a formal comparison. |
| So the rules were bad. | Consequently, the guidelines had a "serious deficiency." | It shows a direct cause-and-effect link. |
| Also, TikTok used tools. | Furthermore, TikTok activated its protocol. | It adds a strong, supporting argument. |
🛠️ How to Use These 'Power Words'
1. The Contrast Marker: "In contrast" Don't just say something is different. Use In contrast at the start of a sentence to tell the reader: "I am now switching to a completely opposite example."
- Example from text: YouTube failed In contrast, TikTok succeeded.
2. The Result Marker: "Consequently" Use this when the second fact is a direct result of the first. It is a formal replacement for so.
- Example from text: The video didn't break rules Consequently, the system is deficient.
3. The Addition Marker: "Furthermore" When you have already given one piece of evidence and want to add a second, even stronger point, use Furthermore. It is the "professional" version of and or also.
- Example from text: TikTok uses tools Furthermore, they have a crisis protocol.
💡 Pro-Tip for Fluency
B2 speakers don't just use these words; they follow them with a comma.
Connector , Full Sentence.
(e.g., Furthermore, the data shows a gap.)
Vocabulary Learning
Examination of Social Media Content Moderation Protocols by the Royal Commission into Antisemitism and Social Cohesion
關於反猶太主義與社會凝聚力的皇家委員會審查社交媒體內容審核協定
Introduction
The Royal Commission into Antisemitism and Social Cohesion has conducted hearings regarding the efficacy of content moderation policies employed by major digital platforms in response to antisemitic discourse and disinformation.
關於反猶太主義與社會凝聚力的皇家委員會已就各大數位平台在應對反猶太言論與虛假訊息時所採用的內容審核政策之成效，舉行了聽證會。
Main Body
The proceedings highlighted a significant divergence in the operational methodologies of the examined entities. YouTube's adherence to its internal guidelines was scrutinized following the retention of a video alleging that a survivor of the Bondi Beach massacre, Arsen Ostrovsky, was a 'crisis actor' involved in a 'false flag' operation. Despite the presence of claims regarding the use of theatrical makeup and the victim's alleged status as an intelligence asset, Rachel Lord, Google's head of government affairs in Australia, asserted that the content was deemed non-violative upon senior-level review. This determination prompted Counsel Assisting Richard Lancaster to characterize the platform's guidelines as possessing a 'serious deficiency.'
程序凸顯了受查機構在操作方法上的顯著分歧。YouTube 在處理一段指稱邦代海灘大屠殺倖存者 Arsen Ostrovsky 為「危機演員」並參與「偽旗行動」的影片時，其對內部指引的遵循情況受到質詢。儘管影片中聲稱使用了舞台化妝，且指稱受害者為情報資產，但 Google 澳洲政府事務主管 Rachel Lord 主張，經高層審查後，該內容被判定為未違規。這一決定促使助理法律顧問 Richard Lancaster 將該平台的指引定性為存在「嚴重缺陷」。
Conversely, TikTok presented a model of proactive containment. Global head of policy Zachary Hecht testified that the platform utilizes automated moderation to assess all uploads, claiming a 98% proactive removal rate for violating content in Australia during the first quarter of the year. The platform's 'crisis response protocol' was reportedly activated within one hour of the Bondi Beach incident, involving coordination with the eSafety Commission and law enforcement. However, the commission identified a systemic vulnerability wherein human moderators occasionally overturned correct automated decisions to remove hate speech, necessitating further escalation by external organizations such as the Dor Foundation to achieve content deletion.
相反地，TikTok 展示了一種主動遏制的模式。全球政策主管 Zachary Hecht 證稱，該平台利用自動化審核來評估所有上傳內容，聲稱今年第一季在澳洲有 98% 的違規內容被主動刪除。據報導，該平台的「危機應對協定」在邦代海灘事件發生一小時內即被啟動，其中涉及與 eSafety 委員會及執法部門的協調。然而，委員會發現了一個系統性漏洞，即人工審核員偶爾會推翻正確的自動化刪除仇恨言論決定，導致需要由 Dor Foundation 等外部組織進一步促請才能達成內容刪除。
Comparative data provided by the Online Hate Prevention Institute suggests a hierarchy of responsiveness among platforms. Analysis indicates that TikTok demonstrated a higher rate of removal for reported harmful content (64%) relative to Meta (54%) and X (24%).
由網絡仇恨防治研究所提供的比較數據顯示，各平台之間存在反應能力的階層差異。分析指出，TikTok 對於被舉報有害內容的刪除率（64%）高於 Meta（54%）及 X（24%）。
Conclusion
The commission continues its inquiry into the role of digital platforms in facilitating online hate, with further testimonies scheduled.
委員會將繼續調查數位平台在助長網路仇恨中所扮演的角色，並已安排進一步的證詞聆訊。
Vocabulary Learning
The Architecture of 'Institutional Detachment'
To transition from B2 to C2, a student must move beyond simple 'formal' language and master Nominalization and Syntactic Distancing. The provided text is a masterclass in Institutional Detachment—the ability to describe volatile, emotional, or controversial events using a linguistic veil of objectivity.
⚡ The Pivot: From Action to Entity
Observe how the text avoids emotive verbs in favor of complex noun phrases. This shifts the focus from who is doing what to what is occurring as a systemic process.
- B2 Approach: "YouTube kept a video that lied about a victim, and a lawyer said their rules were bad."
- C2 Execution: "The retention of a video... prompted Counsel Assisting Richard Lancaster to characterize the platform's guidelines as possessing a 'serious deficiency.'"
Analysis: Note the transformation of the verb retain into the noun retention. This creates a "conceptual object" that can be analyzed clinically. The phrase "possessing a serious deficiency" replaces "are bad," upgrading a subjective judgment to a formal diagnostic statement.
🔍 Lexical Precision & The 'Hedge'
C2 mastery requires the use of precise, high-register terminology to describe systemic failures without sounding anecdotal.
*"...identified a systemic vulnerability wherein human moderators occasionally overturned correct automated decisions..."
- Systemic Vulnerability: Rather than saying "a mistake in the system," the author uses a term from risk management. This implies the flaw is inherent to the design, not just a random error.
- Overturned: A legalistic verb that elevates the act of deleting/changing a decision to a formal procedural action.
🛠️ Structural Sophistication: The Contrastive Frame
The text employs a binary structural device: The Divergence Framework.
Instead of saying "YouTube did this, but TikTok did that," the author introduces the concept of "a significant divergence in the operational methodologies." By establishing this theoretical framework first, the subsequent examples (YouTube's failure vs. TikTok's proactivity) serve as evidence for a thesis rather than just a list of facts. This is the hallmark of academic C2 writing: Framework Evidence Conclusion.