📰 ReutersReutersCodex · 实盘不做把握 5/10
北京时间·🕒 发布 01:01:15时区推断·⟳ 抓取 01:01:48·⚡ 决策 08-01 01:31:20 +29m32s·87.8 天前
美国法官允许Reddit对Perplexity数据抓取诉讼的核心主张继续推进
裁定保留Reddit关于Perplexity绕过保护措施抓取内容用于AI训练的核心主张，方向上支持Reddit的数据授权与内容控制叙事。
📄 原文Reuters↗ 原文
A Manhattan federal judge on Friday rejected most of Perplexity AI's bid to dismiss a lawsuit brought by Reddit (RDDT.N…
不做 Reddit（RDDT.US）
裁定只是诉讼程序性推进，未披露赔偿或授权增量，数据授权叙事近期已有扩散；同时stock_quote两次504且price_snapshot超时，未取得真实现价，不能给可执行ENTER。
展开详情(成交明细 / 研究档案 / 原文全文 / 决策记录链)
研究档案
🔎 新鲜度:最早出现 2026-07-31 · 扩散 medium · price-in 风险 medium ↗
消息传播链
- Reddit此前已起诉Perplexity及数据抓取公司，当前裁定不是诉讼首次曝光。
- Reuters报道曼哈顿联邦法官驳回Perplexity大部分撤诉请求，允许Reddit核心抓取和合谋主张推进。市场反应:行情工具超时，未核到RDDT.US当时实盘反应。
关键理由
- 裁定保留Reddit关于Perplexity绕过保护措施抓取内容用于AI训练的核心主张，方向上支持Reddit的数据授权与内容控制叙事。
- 但这只是驳回撤诉动议后的程序性胜利，Reuters称损害赔偿金额未披露，短期收入、利润或现金流增量无法量化。
- 案例库显示Reddit数据授权和Google/OpenAI授权议题近期已多次被讨论，本条新增事实不足以单独重估RDDT.US。
主要风险
- 诉讼仍需继续审理，Perplexity已明确抗辩，最终赔偿或禁令结果不确定。
- stock_quote与price_snapshot服务超时，未取得RDDT.US真实现价，按ENTER纪律不能给可执行入场。
反向证据
- 做多前查反向 · 来源 kb_search · 影响 none
内部KB检索实际调用但服务504，未能补到付费库反证或量级证据。
检索:Reddit Perplexity motion dismiss Reddit lawsuit July 2026 - 做多前查反向 · 来源 search_case_library · 日期 2026-07-23 · 影响 weakens
案例库显示Reddit数据授权、Google访问限制和约6000万美元年授权关系近期已多次被处理，说明相关AI数据授权叙事已有扩散。
检索:RDDT.US历史决策案例库 - 做多前查反向 · 来源 web · 日期 2026-07-31 · 影响 weakens
Reuters同时披露法官驳回了部分次要主张，且赔偿金额未披露，未形成可计算短线增量。
检索:Reuters Reddit Perplexity lawsuit裁定正文 原文 ↗
工具调用 · 历史记录,结果未知(5 种)
- search_processed_events · 结果未知
- stock_quote · 结果未知
- kb_search · 结果未知
- search_case_library · 结果未知
- price_snapshot · 结果未知
原文全文
展开原文 · 3,199 字
A Manhattan federal judge on Friday rejected most of Perplexity AI's bid to dismiss a lawsuit brought by Reddit (RDDT.N) accusing it of violating U.S. copyright law by scraping the online discussion platform's data to train Perplexity's AI-powered search engine.
U.S. District Judge Paul Engelmayer said in his decision that Reddit could continue pressing its claims that Perplexity and three data scrapers unlawfully circumvented protective measures to steal content for AI training.
Among other things, Engelmayer said that Reddit had standing to sue for Perplexity's alleged misuse of its users' content.
"Reddit’s suit claims the right to control access to public web pages it doesn't own, using a security tool it didn't build, on behalf of users it hasn't asked," a Perplexity spokesperson said. "We’re going to defend the open internet, and we’re going to win."
"Today’s ruling brings us one step closer to holding bad actors accountable," a Reddit spokesperson said. "Reddit supports responsible access to public content, but we oppose companies that bypass our protections, ignore our rules, and profit off our communities without permission."
The case is one of many filed by content owners including authors, music labels and news outlets against tech companies over the alleged misuse of copyrighted material to train AI systems. Reddit filed a separate data-scraping lawsuit against Anthropic last year that is still ongoing in California state court.
Reddit, which features thousands of interest-based "subreddit" web communities, said in the lawsuit against Perplexity that the platform is the most commonly cited source for AI-generated answers to user questions. It has licensed its content to Google, OpenAI and others for their AI training.
Reddit said that Lithuania-based Oxylabs, Russia-based AWMProxy and Texas-based SerpApi -- which it also sued -- scraped Reddit data from billions of search results without permission and that Perplexity, which does not have a license to Reddit content, worked with at least one of them to obtain its material.
SerpApi attorney Jeff Homrig of Weil Gotshal & Manges said in a statement that the company "accesses public search results, not Reddit’s platform, and public information does not become protected because a platform wants to charge for it." Spokespeople for Oxylabs did not immediately respond to a request for comment, and AWMProxy could not immediately be reached for comment.
Reddit asked the court for unspecified monetary damages and an order blocking Perplexity from using its data. Perplexity has denied the allegations.
Engelmayer on Friday dismissed some of Reddit's secondary claims, but advanced its claims that Perplexity unlawfully scraped its data and engaged in a conspiracy with the data-scraping companies.
The case is Reddit Inc v. SerpApi LLC, U.S. District Court for the Southern District of New York, No. 1:25-cv-08736.
For Reddit: Reid Bolton, Matthew Ford and William Gohl of Bartlit Beck
For Perplexity: Eric MacMichael, Sharif Jacob, Benjamin Rothstein and Christina Lee of Keker Van Nest & Peters
Read more:
Reddit sues Perplexity for scraping data to train AI system
U.S. District Judge Paul Engelmayer said in his decision that Reddit could continue pressing its claims that Perplexity and three data scrapers unlawfully circumvented protective measures to steal content for AI training.
Among other things, Engelmayer said that Reddit had standing to sue for Perplexity's alleged misuse of its users' content.
"Reddit’s suit claims the right to control access to public web pages it doesn't own, using a security tool it didn't build, on behalf of users it hasn't asked," a Perplexity spokesperson said. "We’re going to defend the open internet, and we’re going to win."
"Today’s ruling brings us one step closer to holding bad actors accountable," a Reddit spokesperson said. "Reddit supports responsible access to public content, but we oppose companies that bypass our protections, ignore our rules, and profit off our communities without permission."
The case is one of many filed by content owners including authors, music labels and news outlets against tech companies over the alleged misuse of copyrighted material to train AI systems. Reddit filed a separate data-scraping lawsuit against Anthropic last year that is still ongoing in California state court.
Reddit, which features thousands of interest-based "subreddit" web communities, said in the lawsuit against Perplexity that the platform is the most commonly cited source for AI-generated answers to user questions. It has licensed its content to Google, OpenAI and others for their AI training.
Reddit said that Lithuania-based Oxylabs, Russia-based AWMProxy and Texas-based SerpApi -- which it also sued -- scraped Reddit data from billions of search results without permission and that Perplexity, which does not have a license to Reddit content, worked with at least one of them to obtain its material.
SerpApi attorney Jeff Homrig of Weil Gotshal & Manges said in a statement that the company "accesses public search results, not Reddit’s platform, and public information does not become protected because a platform wants to charge for it." Spokespeople for Oxylabs did not immediately respond to a request for comment, and AWMProxy could not immediately be reached for comment.
Reddit asked the court for unspecified monetary damages and an order blocking Perplexity from using its data. Perplexity has denied the allegations.
Engelmayer on Friday dismissed some of Reddit's secondary claims, but advanced its claims that Perplexity unlawfully scraped its data and engaged in a conspiracy with the data-scraping companies.
The case is Reddit Inc v. SerpApi LLC, U.S. District Court for the Southern District of New York, No. 1:25-cv-08736.
For Reddit: Reid Bolton, Matthew Ford and William Gohl of Bartlit Beck
For Perplexity: Eric MacMichael, Sharif Jacob, Benjamin Rothstein and Christina Lee of Keker Van Nest & Peters
Read more:
Reddit sues Perplexity for scraping data to train AI system
决策记录链 · 只追加
每次领取产生一条记录,重新决策追加一条并以 supersedes 指向被替代的记录(ADR 0014)。
这条决策早于决策账本(2026-10-08 起逐条追加),没有追加记录。