Anthropic 就大規模人工智慧版權訴訟達成和解
Anthropic Settles Massive AI Copyright Lawsuit
Updated at: August 14, 2026 at 01:00 AM
在2026年7月，當Anthropic與一群作者達成了一項歷史性的15億美元和解協議時，AI監管的格局發生了重大轉變。
In July 2026, the landscape of AI regulation shifted significantly when Anthropic reached a historic $1.5 billion settlement with a group of authors.
Anthropic訴訟案，該案指控這家AI開發商利用盜版書籍來訓練其Claude模型。
Anthropic, which accused the AI developer of using pirated books to train its Claude models.
這場衝突始於2024年，當時作者們指控Anthropic從影子圖書館中獲取了數十萬部受版權保護的著作。
The conflict began in 2024 when authors alleged that Anthropic obtained hundreds of thousands of copyrighted works from shadow libraries.
威廉·奧爾薩普法官在2025年的一項關鍵裁決有助於形成這一結果：他裁定，雖然AI訓練的整體行為可能符合「合理使用」，但從非法資料庫下載盜版材料則不然。
A crucial 2025 ruling by Judge William Alsup helped shape the outcome: he determined that while the general act of AI training may qualify as 'fair use,' downloading pirated material from illegal databases does not.
面臨超過700億美元潛在損害賠償的威脅，Anthropic選擇了和解。
Facing potential damages exceeding $70 billion, Anthropic chose to settle.
雖然此項和解免除了Anthropic在數據獲取方面過去的責任，但它仍然是一個微妙的判例。
While this settlement clears Anthropic of past liability regarding data acquisition, it remains a nuanced precedent.
它既沒有保護該公司免受未來關於AI生成輸出內容的訴訟，也沒有最終解決圍繞AI對受版權保護材料進行訓練是否在法律上被允許的更廣泛爭議。
It does not shield the company from future litigation over AI-generated outputs, nor does it definitively settle the broader debate surrounding whether AI training on copyrighted material is legally permissible.
