News Companies Sue OpenAI
News Companies Sue OpenAI
新聞公司起訴 OpenAI
Introduction
Some news companies in the US are angry with OpenAI. They asked a court to punish the company.
美國一些新聞公司對 OpenAI 表示不滿，他們請求法院對該公司進行懲罰。
Main Body
News companies say OpenAI stole their stories to train AI. OpenAI said they could not search their data. But a worker said this is not true. OpenAI did search the data.
新聞公司表示 OpenAI 盜用他們的報導來訓練 AI。OpenAI 聲稱他們沒有搜尋其數據。但一名員工表示這並非事實，OpenAI 的確搜尋了該數據。
The news companies say OpenAI deleted many files. They also say OpenAI hid important information in chat logs. They want the judge to punish OpenAI for these lies.
這些新聞公司表示 OpenAI 刪除了許多檔案。他們還指稱 OpenAI 在對話紀錄中隱瞞了重要資訊。他們希望法官能針對這些謊言懲罰 OpenAI。
AI is changing how people read news. Some news companies make deals with AI. Other companies go to court. OpenAI says they did nothing wrong. They say they want to protect user privacy.
AI 正在改變人們閱讀新聞的方式。一些新聞公司與 AI 達成協議，而其他公司則選擇走法律途徑。OpenAI 則表示他們沒有做錯任何事，並稱其目的是為了保護用戶隱私。
Conclusion
Now the judge must decide if OpenAI lied and if they must be punished.
現在法官必須決定 OpenAI 是否說謊，以及是否應受到懲罰。
Vocabulary Learning
⚡ Action words in the past
In this story, things already happened. We change the words to show it is the past.
The Pattern: Word + -ed = Past time
- ask asked
- punish punished
- search searched
- delete deleted
⚠️ The 'Rule Breakers' Some words do not use -ed. You must remember these alone:
- say said
- steal stole
- do did
- hide hid
Example from text: "OpenAI said they could not search... But a worker said this is not true."
Vocabulary Learning
Lawsuit Over Alleged Evidence Misconduct by OpenAI in Copyright Case
OpenAI 涉嫌在版權案件中操縱證據而面臨訴訟
Introduction
Several US media companies, including The New York Times and the New York Daily News, have asked a federal court in Manhattan to punish OpenAI. This request follows claims that the company hid evidence and gave false information during the legal process.
包括《紐約時報》與《紐約日報》在內的幾家美國媒體公司，已要求曼哈頓聯邦法院懲罰 OpenAI。此請求是基於該公司在法律程序中隱瞞證據並提供錯誤資訊的指控。
Main Body
The legal battle began with a 2023 lawsuit claiming that OpenAI and Microsoft used copyrighted news articles to train AI models without permission. The current issue is that the media companies believe OpenAI lied about its ability to search through training data and user chat logs. Although OpenAI previously claimed these searches were technically impossible and would violate user privacy, a statement from an engineer suggests that the company had actually checked for copyrighted material before the lawsuit started.
這場法律戰始於 2023 年的一項訴訟，指控 OpenAI 與微軟在未經許可的情況下，使用受版權保護的新聞文章來訓練 AI 模型。目前的爭議在於，媒體公司認為 OpenAI 在是否能搜尋訓練數據和使用者聊天記錄的能力上撒謊。儘管 OpenAI 先前聲稱這些搜尋在技術上是不可能的，且會違反使用者隱私，但一名工程師的聲明顯示，該公司在訴訟開始前實際上已檢查過版權材料。
Furthermore, the plaintiffs claim that OpenAI used a secret system called 'Project Giraffe' to monitor if the AI was repeating copyrighted text. They also argue that OpenAI deleted billions of records and provided 20 million chat logs that were useless because too much information had been removed. Consequently, the media companies are asking the court to reject these logs as evidence and assume that the missing data would have proven that OpenAI copied their content.
此外，原告指控 OpenAI 使用了一個名為「長頸鹿計畫」(Project Giraffe) 的秘密系統，用以監控 AI 是否重複受版權保護的文本。他們還主張 OpenAI 刪除了數十億條記錄，並提供了 2,000 萬份毫無用處的聊天記錄，因為過多資訊已被刪除。因此，媒體公司要求法院拒絕將這些記錄作為證據，並推定缺失的數據將能證明 OpenAI 抄襲了他們的內容。
This case is part of a larger trend where AI-generated summaries are reducing web traffic and ad revenue for traditional news. While some news outlets have signed deals to let AI companies use their work, others are fighting in court. OpenAI has denied these accusations, asserting that the media companies are simply trying to invade user privacy because their copyright claims are becoming weaker.
本案是更大趨勢的一部分，即 AI 生成的摘要正在減少傳統新聞的網頁流量與廣告收入。雖然部分新聞機構已簽署協議允許 AI 公司使用其作品，但其他機構則選擇在法庭上對抗。OpenAI 否認了這些指控，堅稱媒體公司僅僅是試圖侵犯使用者隱私，因為他們的版權主張正變得越來越乏力。
Conclusion
The court must now decide if OpenAI behaved dishonestly during the discovery process and if the requested penalties are necessary.
法院現在必須決定 OpenAI 在證據披露過程中是否表現不誠實，以及所請求的處罰是否必要。
Vocabulary Learning
⚡ The Power of "Connectors" (A2 → B2 Jump)
At an A2 level, you likely use simple words like and, but, or because. To reach B2, you need to use Logical Transitions. These are words that act like road signs, telling the reader exactly how the next idea relates to the previous one.
🔍 Analysis from the Text
Look at these three specific words used in the article. They change the entire "feel" of the writing from a simple story to a professional argument:
-
Furthermore Used to add a second, stronger point.
- A2 style: "They also said OpenAI used a secret system."
- B2 style: "Furthermore, the plaintiffs claim that OpenAI used a secret system..."
-
Consequently Used to show a direct result (Cause Effect).
- A2 style: "So, the media companies want the court to reject logs."
- B2 style: "Consequently, the media companies are asking the court to reject these logs..."
-
While Used to compare two opposite situations in one sentence.
- A2 style: "Some outlets signed deals. Other outlets are fighting."
- B2 style: "While some news outlets have signed deals... others are fighting in court."
🛠️ Practical Application Guide
| Instead of... (A2) | Try this... (B2) | Why? |
|---|---|---|
| And / Also | Furthermore / Moreover | It sounds more academic and persuasive. |
| So | Consequently / Therefore | It proves a logical link between two facts. |
| But | While / Whereas | It allows you to balance two ideas together. |
Pro Tip: Don't just use these at the start of a sentence. Try placing them after the subject for a more sophisticated flow: "The companies, consequently, asked the court for a penalty."
Vocabulary Learning
Litigation Regarding Alleged Discovery Misconduct by OpenAI in Copyright Infringement Proceedings
關於 OpenAI 在版權侵權訴訟中涉嫌證據披露不當的法律爭議
Introduction
Several United States media organizations, including The New York Times and the New York Daily News, have petitioned a federal court in Manhattan to impose sanctions on OpenAI following allegations of evidence suppression and misrepresentation.
包括《紐約時報》與《紐約每日新聞》在內的數家美國媒體機構，已向曼哈頓聯邦法院申請對 OpenAI 採取制裁，理由是其涉嫌隱瞞證據與做出虛假陳述。
Main Body
The current legal dispute originates from a 2023 lawsuit asserting that OpenAI and Microsoft utilized copyrighted journalistic content to train generative artificial intelligence models without authorization. Central to the current motion is the contention that OpenAI provided inaccurate testimony regarding its technical capacity to query training datasets and user chat logs. While the defendant previously maintained that such searches were technically prohibitive and would compromise user anonymity, a deposition by data privacy engineer Vinnie Monaco suggests that internal evaluations of copyrighted material had been conducted prior to the initiation of the lawsuit.
目前的法律爭議源於 2023 年的一場訴訟，該訴訟主張 OpenAI 與微軟在未經授權的情況下，利用受版權保護的新聞內容來訓練生成式人工智慧模型。本次動議的核心爭議在於，原告認為 OpenAI 針對其查詢訓練數據集與用戶對話日誌的技術能力，提供了不準確的證詞。儘管被告先前堅稱此類搜尋在技術上不可行且會危及用戶匿名性，但數據隱私工程師 Vinnie Monaco 的證詞顯示，在訴訟啟動前，內部已對受版權保護的資料進行過評估。
Furthermore, the plaintiffs allege the existence of 'Project Giraffe' and a corresponding 'Bloom' filter designed to monitor output regurgitation, alongside the maintenance of a database containing approximately 78 million de-identified conversations. The plaintiffs assert that OpenAI's subsequent submission of 20 million chat logs was rendered non-functional through excessive redaction and that billions of records were deleted in contravention of court-mandated preservation orders. Consequently, the plaintiffs seek judicial sanctions, including the disqualification of the provided logs as evidence and a legal presumption that the logs would have demonstrated substantial content reproduction.
此外，原告指稱存在一個名為「長頸鹿計畫」（Project Giraffe）的專案以及相對應的「Bloom」過濾器，旨在監控輸出內容的重複率，同時維護一個包含約 7,800 萬筆去識別化對話的資料庫。原告主張，OpenAI 隨後提交的 2,000 萬筆對話日誌因過度遮蔽而失去功能，且數十億筆記錄在違反法院保存令的情況下被刪除。因此，原告尋求司法制裁，包括取消所提供日誌的證據資格，並在法律上推定該日誌將證明存在實質性的內容複製。
This litigation occurs within a broader systemic shift where AI-generated summaries are impacting traditional news traffic and advertising revenue. While some media entities have entered into licensing agreements with AI developers, others continue to seek judicial remedies. OpenAI has formally denied these allegations, characterizing the plaintiffs' requests as an attempt to infringe upon user privacy in response to a perceived weakening of the underlying copyright claims.
此次訴訟發生在更廣泛的系統性轉型背景下，即 AI 生成的摘要正影響著傳統新聞的流量與廣告收入。雖然部分媒體實體已與 AI 開發商簽署授權協議，但其他實體仍繼續尋求司法救濟。OpenAI 已正式否認這些指控，將原告的請求描述為在感知到版權主張削弱後，試圖侵犯用戶隱私的行為。
Conclusion
The court must now determine whether OpenAI engaged in discovery misconduct and if the requested sanctions are warranted.
法院現在必須判定 OpenAI 是否涉及證據披露不當，以及所請求的制裁是否合理。
Vocabulary Learning
The Architecture of 'Legalistic Precision' and Nominalization
To transition from B2 to C2, a learner must move beyond describing events to constructing institutional reality. The provided text is a masterclass in Nominalization—the process of turning verbs (actions) into nouns (concepts). This is the linguistic hallmark of high-level legal and academic discourse, as it shifts the focus from who is doing what to the abstract state of affairs.
◈ The Shift: Action Entity
Observe how the text avoids simple active verbs in favor of complex noun phrases:
- B2 approach: OpenAI might have hidden evidence and lied about it.
- C2 approach: ...allegations of evidence suppression and misrepresentation.
By transforming the verbs suppress and misrepresent into the nouns suppression and misrepresentation, the author creates a 'frozen' concept that can be manipulated as a legal object. This allows for the introduction of high-precision modifiers (e.g., alleged, systemic) without cluttering the sentence structure.
◈ Syntactic Density: The 'Noun-Heavy' Chain
C2 proficiency is marked by the ability to handle dense noun clusters. Consider this segment:
"...a legal presumption that the logs would have demonstrated substantial content reproduction."
Breakdown of the cognitive load:
- Legal presumption (The conceptual framework)
- Substantial content reproduction (The technical phenomenon)
In a B2 context, this would be: "The court should assume the logs showed that a lot of content was copied." While correct, the B2 version is narrative; the C2 version is propositional. It treats the entire idea of 'copying content' as a single, stable entity (content reproduction).
◈ Lexical Nuance: The 'Hedging' Spectrum
Note the strategic use of verbs that distance the author from the claims, a critical skill for C2 academic writing:
- Asserting Contending Characterizing
While a B2 student might use "say" or "claim," the C2 writer chooses "contention" to signal a formal legal argument and "characterizing" to suggest a strategic framing of the truth. This is not just vocabulary; it is rhetorical positioning.