OpenAI 行動
287 人已加入OpenAI 行動全部成員
已驗證電子郵件並加入這個行動的帳號數；不代表法律資格已經查核。
給認為自己的作品可能被 OpenAI 產品（例如 ChatGPT）使用的人。找到相似經歷的人，了解已知情況，知道自己可以先做什麼。
接下來怎麼做
選擇你的作品類型，看對應的指南。這些分類只用來篩選內容，不會產生分組人數。
指南：你可以自己做什麼
- 材料清单：E 编号是什么意思
- 记录底稿：11 份可以照抄的空白表
- SOP02｜先保住眼前的现场
- SOP03｜保存完整输入、输出和相关会话
- SOP04｜制作原作与输出对照
- SOP05｜生成指纹、记保存过程并安全备份
- SOP06｜整理既有发表、授权和权利记录
- 还不确定属于哪一类：从这里开始
另有其他語言版本 (8)
- The material list: what the E numbers mean (英文)
- Record templates: ten blank notes you can copy (英文)
- SOP02 | Preserve what is in front of you first (英文)
- SOP03 | Save the full input, output and related conversation (英文)
- SOP04 | Build a comparison between the work and the output (英文)
- SOP05 | Produce fingerprints, record the process and back it up safely (英文)
- SOP06 | Organise existing publication, licence and rights records (英文)
- Not sure which category: start here (英文)
各地的規則
和你有同樣疑問的人不止一個。這些頁面寫的是你所在地的規則實際怎麼說，每條都附來源。是背景資料，不是法律建議。
- 加拿大：规则概览
- 中国内地：规则概览
- 西班牙：规则概览
- 欧盟：规则概览
- 印度：规则概览
- 墨西哥：规则概览
- 美国：规则概览
- Canada: what the rules say (英文)
- China (mainland): what the rules say (英文)
- Spain: what the rules say (英文)
- European Union: what the rules say (英文)
- India: what the rules say (英文)
- Mexico: what the rules say (英文)
- United States: what the rules say (英文)
案件與進展
- AI Copyright Deal: Albanese and Taylor Move Toward a Licence-Free Model
- Proposed class action accuses OpenAI of using minors' ChatGPT data to train models
- OpenAI accused of taking minors' data without parental consent to train AI models
- Australian panel urges PM to tighten AI safeguards after OpenAI breaches; calls for creator-focused copyright review
- UK privacy watchdog secures AI company commitments on how personal data is used to train AI models
- Leaked Australian proposal would allow AI training opt-out, with no compensation under one option
- OpenAI publishes crawler controls: GPTBot training opt-out via robots.txt
- U.S. Justice Department files statement of interest backing fair use for AI training in New York Times case
相關參考（與 OpenAI 無關）
- AI Lawsuit Tracker counts 248 cases against 76 companies; copyright dominates — Named as the most-sued company in the tracker (55 cases).
- Anthropic says new safeguards cut AI model-distillation attacks by 90–99% — Named as one of the leading labs that share Anthropic's concern about distillation letting rivals copy model capabilities.
- EFF guide: how to opt out of AI training on ChatGPT, Gemini, Claude, Copilot and other major platforms — Guide links to ChatGPT's own documentation for opting out of training on conversations.
- European Commission opens consultation on copyright and AI, asking whether rules on protected content in AI need strengthening — Major general-purpose AI provider placing models on the EU market; the consultation seeks input from generative AI providers and reviews the copyright framework governing their use of protected content. No single company is the subject.
- Judge lets three of Reddit's claims against Anthropic proceed and dismisses two — OpenAI licensed Reddit data for training; named only for contrast, the dispute is not against OpenAI
討論
瀏覽討論不需要登入；發文需要登入並加入本行動。
已查核律師發布
還沒有已查核律師在這裡發表內容。