Core Insights - Anthropic reported that three organizations systematically created over 24,000 fraudulent accounts, resulting in more than 16 million interactions with Claude, aimed at extracting model capabilities for their own training and optimization [1][7]. Group 1: Distillation Activities - The three distillation actions exhibited highly similar operational methods, utilizing fake accounts and proxy services for large-scale access to evade platform detection [7]. - Anthropic identified these actions through multiple technical evidences, including IP address associations and request metadata, achieving high-confidence attribution [7]. - The attacks primarily targeted Claude's differentiated capabilities in agentic reasoning, tool usage, and code generation [7]. Group 2: DeepSeek Investigation - In the investigation of DeepSeek, Anthropic confirmed that the scale of operations exceeded 150,000 interactions, focusing on multi-task reasoning and sensitive question rephrasing [8]. - DeepSeek's accounts displayed synchronized traffic patterns and payment methods, resembling a "load balancing" feature to enhance throughput and reduce detection risk [8]. - One identified technique involved prompting Claude to "retrace and write out its internal reasoning process," generating large-scale chain-of-thought training data [8]. Group 3: Moonshot AI and MiniMax - For Moonshot AI, Anthropic disclosed over 3.4 million interactions, concentrating on agentic reasoning, programming, and computer vision capabilities [8]. - Moonshot employed hundreds of fraudulent accounts and mixed various access paths to lower the overall recognizability of their actions [8]. - The largest distillation activity was attributed to MiniMax, with over 13 million interactions, focusing on agentic programming capabilities and tool orchestration [8]. - Anthropic was able to observe the entire process of a distillation attack from data generation to model release, as MiniMax adjusted its strategy shortly after the release of a new model [8]. Group 4: Security Measures - Anthropic stated that the findings have been used to enhance the platform's security and abuse detection mechanisms, although further details on actions taken were not disclosed [9]. - As of the report's publication, DeepSeek, Moonshot AI, and MiniMax had not responded to the situation [9].

证券研究报告 2026.02.11 人工智能十年展望（二十七）：越过"遗忘" 的边界，模型记忆的三层架构与产业机遇 SAC 执证编号：S0080518070011 SFC CE Ref：BOP246 于钟海 分析员 韩蕊 分析员 王之昊 分析员 SAC 执证编号：S0080523070010 SFC CE Ref：BXD683 rui.han@cicc.com.cn SAC 执证编号：S0080522050001 SFC CE Ref：BSS168 zhihao3.wang@cicc.com.cn 纵轴：相对值（%） 88 100 112 124 136 148 2025-02 2025-05 2025-08 2025-10 2026-01 沪深300 中金软件及服务 投资建议 大模型的演进史，本质上是一部与"遗忘"抗争的历史。当我们惊 叹于模型的推理能力时，往往忽视了一个重要短板：在缺乏记忆留 存的架构下，模型每一次对历史信息的处理，本质上都是一次昂贵 的"重复计算"。这种以高昂算力对抗遗忘的粗放模式，正面临着显 存墙与上下文窗口的物理极限。我们认为，2026 年及之后的AI Infra主战场将增加"模型记 ...

Anthropic今天指控DeepSeek、MiniMax和Moonshot蒸馏Claude的能力这波操作，属实是把贼喊捉贼玩出 了新高度。 我总结成两句话——美国AI商业上卷不动了，就开始掀桌子；技术上又不自信了，就开始搞政治投 机。 1 先把蒸馏这个听起来高大上的技术名词祛魅。说白了，就是AI大模型公司甲付费买了AI大模型公司乙 的API服务，通过格式化提问的方式，拿到了合法的输出结果，然后用这些结果来训练自己的大模型。 先讲业务合理性，在现有的AI竞争环境里，蒸馏本来就是一个司空见惯的手段。现在哪个大模型公司 敢说自己没有蒸馏过别人的大模型？这一直是行业里心照不宣的进化手段和业务规则。 再讲商业，Anthropic自己开门做着API的生意，收着真金白银的调用费，现在我中国AI公司一手交钱， 一手拿货。货（数据）到了我手里，我是拿来做PPT、写代码，还是拿来喂给我的模型做教案，关你屁 事？ 大家都在一个池子里摸鱼，现在中国公司摸了条大鱼，Anthropic就跳脚了。这跟开了家自助餐厅，嫌 客人吃回本了就报警吵着要抓人有什么区别？ 无论在业务还是商业上，中国AI公司根本没有原罪。 2 但更讽刺的是，Anth ...

4、美称中国一人工智能企业违反美出口管制，外交部：中方已多次表明原则立场 1、特朗普考虑征收新的国家安全关税 美国最高法院上周的一项裁决宣布总统特朗普第二任期的多项征税无效，现在特朗普政府考虑对六 个行业征收新的国家安全关税。据知情人士透露，考虑征收的新关税可能涵盖大型电池、铸铁和铁配 件、塑料管道、工业化学品以及电网和电信设备等行业。这些关税将根据《1962年贸易扩展法》(Trade Expansion Act of 1962)第232条征收，该条款赋予总统基于国家安全风险征收关税的广泛权力。新的第 232条关税将独立于特朗普自最高法院周五上午驳回其多项关税以来已宣布的其他税项。已宣布的关税 包括一项新的15%关税，可维持五个月，以及计划在该期限后征收的多项关税，后者将根据《贸易法》 (Trade Act)第301条发布。 2、美媒：特朗普军方最高顾问警告袭击伊朗风险 据Axios援引两名消息人士透露，美军参谋长联席会议主席丹・凯恩将军已向特朗普总统及高级官 员建议，对伊朗发动军事行动可能存在重大风险，尤其是可能陷入长期冲突。 特朗普政府高层正就是 如何应对伊朗对峙、以及不同方案将带来何种后果展开激烈争论。 ...

1月30日，中共中央政治局就前瞻布局和发展未来产业进行第二十四次集体学习。培育发展未来产业， 对于抢占科技和产业制高点、把握发展主动权，对于发展新质生产力、建设现代化产业体系，对于提高 人民生活品质、促进人的全面发展和社会全面进步，都具有重要意义。多位专家将深入解读谋划和布局 未来产业的关键路径和有效实践。 1月30日下午，中共中央政治局第二十四次集体学习聚焦前瞻布局和发展未来产业，释放出强烈信号： 未来产业在即将到来的"十五五"工作中将占据极端重要地位。如何建立高水平的"未来产业瞭望站"，从 海量创新主体中精准识别真正发挥关键作用的"企业主体"，成为赋能未来产业发展的核心抓手与关键举 措。 作为驱动产业创新变革的核心力量，创造型幸福企业在产业演进中扮演着至关重要的多重角色。它们是 产业技术攻坚者，主动对接国家重大科技与未来产业战略发展需求，长期聚焦产业核心"卡脖子"技术， 以底层核心技术突破撬动产业变革。这类企业致力于攻克最复杂、最关键的硬骨头，它们是产业体系的 压舱石，决定着产业的技术天花板、成本结构与竞争格局，真正实现以技术突破定义产业未来，甚至催 生全新的产业形态。他们还是产业生态重构者，以自身为枢纽 ...

AIPress.com.cn报道 2月24日消息，Anthropic在周一发布的博客文章中称，DeepSeek、Moonshot AI和MiniMax通过约2.4万个虚 假账户，与Claude进行了超过1600万次交互，从而提取模型能力，用于训练和优化自家模型。突发！ Anthropic声称被三家中国AI公司蒸馏 随后，马斯克在社交平台X上发文称，Anthropic"曾大规模窃取训练数据"，并声称其曾因此支付数十亿美 元和解金。他还附上"X社区注释"的截图作为佐证。 在回应网友的质疑时，马斯克直言Anthropic这样的做法是自鸣得意、伪善、虚伪的。 目前，Anthropic尚未就马斯克的相关指控作出进一步回应。 ...

人民财讯2月24日电，2月24日，外交部发言人毛宁主持例行记者会。有记者提问，据一位特朗普政府高 级官员称，DeepSeek的AI模型据说是使用英伟达最先进的AI芯片进行训练的。这可能构成对美国出口 管制的违反，美方认为DeepSeek需要将相关设备移除。请问这一说法是否属实？ 转自：证券时报 毛宁表示，不了解你提到的具体情况。关于美国输华芯片问题，中方已经多次表明了原则立场。 ...

Core Viewpoint - Anthropic has accused three Chinese AI companies—DeepSeek, Moonshot AI, and MiniMax—of large-scale "distillation" of its model Claude, claiming that these companies used over 24,000 fake accounts to interact with Claude approximately 16 million times to extract model capabilities for their own models [1][3][16]. Group 1: Distillation Process - Distillation is a common AI training method where a stronger "teacher model" generates output data to train a "student model," allowing for the replication of some capabilities at a lower cost and parameter scale [2][14]. - The controversy centers on the scale and method of distillation, with Anthropic alleging that the three companies systematically extracted Claude's capabilities through shared payment methods, proxy services, and bulk request structures [3][16]. Group 2: Specific Interactions - DeepSeek is accused of over 150,000 interactions focusing on reasoning and thought chain data; Moonshot AI is reported to have around 3.4 million interactions targeting agent capabilities and tool invocation; MiniMax had the highest number, approximately 13 million interactions, concentrating on agent orchestration and tool usage [3][16]. Group 3: Industry Reactions - Elon Musk criticized Anthropic on social media, suggesting that the company has previously faced controversies regarding training data and implying hypocrisy in their accusations [3][19]. - There are differing opinions within the industry regarding the focus of the controversy, with some arguing that the issue lies not in the distillation technology itself but in the specific implementation methods that may violate service terms or regional restrictions [21][22]. Group 4: Legal and Ethical Considerations - The lack of clear legal standards regarding the ownership of model outputs raises questions about whether the actions of the accused companies constitute normal competition or unfair extraction [23][24]. - The ongoing debate highlights the need for clearer definitions of what constitutes reasonable use versus systematic capability extraction in the context of AI model training [24].

Core Insights - Anthropic has accused three domestic AI companies—DeepSeek, Kimi, and MiniMax—of conducting "distillation attacks" to extract capabilities from its Claude model, claiming these companies used approximately 24,000 accounts and engaged in over 16 million conversations with Claude [1][3]. Group 1: Allegations and Responses - Anthropic's claims include that DeepSeek conducted 150,000 conversations primarily targeting reasoning capabilities, while Kimi engaged in 3.4 million conversations focusing on agent reasoning, tool usage, programming, and computer vision [3]. - MiniMax is reported to have the largest scale of interaction, with over 13 million conversations aimed at agent programming and tool invocation [3]. - The company has described the proxy architecture used by these firms as a "Hydra cluster," which involves managing over 20,000 accounts across various platforms, complicating detection efforts [5]. Group 2: Controversies and Historical Context - Anthropic has faced scrutiny for its own practices, including a secret project named "Project Panama," which involved destructively scanning and digitizing books without consent, processing between 500,000 to 2 million books at a cost of tens of millions of dollars [5]. - The company has also been involved in legal disputes over using pirated e-book libraries for model training, resulting in a $1.5 billion settlement in 2025 [5]. - Public reactions have highlighted the irony of Anthropic's accusations, questioning the ethical implications of its own data sourcing practices [6]. Group 3: Industry Reactions - Elon Musk has commented on the situation, questioning the audacity of these companies in light of Anthropic's own history of data acquisition [7]. - The ongoing debate raises critical questions about intellectual property and ethical standards in AI development [8].

On Monday, xAI CEO Elon Musk escalated his feud after Amazon.com, Inc. (NASDAQ:AMZN) -backed Anthropic accused Chinese firms like DeepSeek of copying its Claude model.Anthropic Alleges ‘Industrial-Scale' AI DistillationAnthropic said that Chinese AI companies, including DeepSeek, Moonshot AI and MiniMax, orchestrated what it described as "industrial-scale" distillation attacks on its Claude model.In a blog post, the company alleged the labs created more than 24,000 fraudulent accounts and generated over 16 ...