Anthropic Accuses Alibaba of Massive AI Model Distillation Attack
Anthropic, maker of the Claude AI, formally accused Chinese tech giant Alibaba of orchestrating the largest known “distillation” attack on its AI models. In a June 10 letter to U.S. senators, Anthropic claimed Alibaba-affiliated operators used nearly 25,000 fraudulent accounts to conduct 28.8 million exchanges with Claude between April and June 2026, aiming to illicitly extract capabilities for Alibaba’s Qwen models. This follows earlier accusations against Chinese AI labs DeepSeek, Moonshot, and MiniMax. The allegations heighten U.S.-China tech tensions amid export controls on advanced semiconductors.
Cross-source coverage
Wire timeline
Alibaba bans Anthropic AI tools for employees after distillation attack accusation
Chinese e-commerce giant Alibaba will ban employees from using Anthropic's AI tools, including Claude Code, starting July 10, placing them on a high-risk software list due to security concerns. This follows Anthropic's June accusation that Alibaba conducted 'the largest known distillation attack' to extract its AI capabilities. Anthropic's terms of service already prohibit Chinese companies from using its models. Alibaba employees must uninstall all Anthropic products and switch to the company's own AI assistant, Qoder. The ban coincides with online backlash in China over hidden code in Claude designed to detect Chinese users. Separately, the Financial Times reported that Ant Group provided employees with corporate Claude accounts accessed via Singapore, while ByteDance reimburses engineers for personal Claude subscriptions through VPNs.
US Top News and AnalysisAlibaba Bans Employees from Using Anthropic's AI Tools Amid Security and Distillation Attack Allegations
Chinese e-commerce giant Alibaba will ban employees from using Anthropic's artificial intelligence tools, including Claude Code, effective July 10, citing back-door security risks. The move follows Anthropic's June accusation that Alibaba orchestrated 'the largest known distillation attack' to illicitly extract its AI capabilities. Anthropic's terms of service explicitly prohibit Chinese companies and other 'adversarial nations' from using its models. Alibaba employees are required to uninstall all Anthropic products and switch to the company's own AI assistant, Qoder. The ban coincides with online backlash in China after reports of hidden code in Claude Code designed to detect Chinese users. Additionally, the Financial Times reported that Chinese fintech Ant and TikTok owner ByteDance have employed workarounds, such as using Singapore-based entities and VPNs, to access Claude. Anthropic is reportedly closing these loopholes. Both Alibaba and Anthropic declined to comment on the matter.
US Top News and AnalysisAlibaba Bans Anthropic's Claude Code Over Alleged China-Detection Backdoor, Widening Rift
Chinese tech giant Alibaba has banned its employees from using Anthropic's Claude Code, effective July 10, after security researchers alleged the AI coding tool contained hidden code to detect users in China or affiliated with Chinese AI labs. The code, reportedly shipped since April 2, checked system timezones and proxy URLs against Chinese domains and AI lab identifiers, exfiltrating data steganographically. Employees were instructed to switch to Alibaba's in-house Qoder platform and uninstall all Anthropic products. The ban escalates a feud that began when Anthropic accused Alibaba's Qwen lab of a massive model distillation attack using 25,000 fraudulent accounts. Anthropic engineer Thariq Shihipar acknowledged the detection mechanism as an experiment to prevent abuse, stating it was removed on July 1. The incident reflects growing U.S.-China AI tensions amid export controls and security concerns.
Latest from Tom's HardwareAlibaba Bans Anthropic's Claude Code Over Alleged Hidden China-Detection Backdoor
Chinese tech giant Alibaba has banned its employees from using Anthropic's Claude Code, effective July 10, after security researchers alleged the AI coding tool contained hidden code designed to detect users in China or affiliated with Chinese AI labs. The code, reportedly shipped since version 2.1.91, checked system timezones and proxy URLs against Chinese domains and AI lab identifiers, exfiltrating data steganographically via system prompts. Alibaba instructed staff to switch to its in-house Qoder platform and uninstall all Anthropic products. The ban escalates a feud that began when Anthropic accused Alibaba's Qwen lab of running a massive model distillation attack using 25,000 fraudulent accounts. Anthropic engineer Thariq Shihipar described the detection mechanism as an experiment to prevent abuse, stating it was removed on July 1. The incident reflects broader U.S.-China AI tensions amid export controls and access restrictions.
Latest from Tom's HardwareAlibaba Bans Anthropic's Claude Code Over Alleged Hidden China-Detection Backdoor
Chinese tech giant Alibaba has banned its employees from using Anthropic's Claude Code, effective July 10, after security researchers alleged the AI coding agent contained hidden code designed to detect Chinese users. The code, reportedly shipped since version 2.1.91, checked system timezones and proxy URLs against Chinese domains and AI lab identifiers, then exfiltrated findings steganographically via system prompts. Anthropic engineer Thariq Shihipar acknowledged the mechanism as an experiment to prevent abuse and distillation, stating the code was removed on July 1. Alibaba instructed staff to switch to its in-house Qoder platform and uninstall all Anthropic products. The ban escalates a feud that began when Anthropic accused Alibaba's Qwen lab of running a massive model distillation attack using 25,000 fraudulent accounts. The incident highlights growing tensions in U.S.-China AI relations amid export controls and access restrictions.
Latest from Tom's HardwareAnthropic Accuses Alibaba of Large-Scale AI Distillation; Impact on BABA Stock
Anthropic, a leading U.S. AI startup, has formally accused Alibaba of orchestrating a massive AI data distillation campaign targeting its Claude models. According to a letter sent to U.S. senators, operators allegedly linked to Alibaba's Qwen AI lab used approximately 25,000 fraudulent accounts to conduct nearly 28.8 million exchanges with Claude models over six weeks from late April to early June 2026. The campaign focused on extracting capabilities in software engineering and agentic reasoning. This comes amid heightened U.S.-China AI tensions, with the Pentagon recently adding Alibaba to a list of firms allegedly supporting China's military (Alibaba denies this and has sued). In response, bipartisan senators are moving to attach an amendment to defense legislation to sanction Chinese firms found improperly accessing U.S. AI model outputs. BABA stock fell nearly 5% on June 25 and over 9% over the past five days. Alibaba has not yet commented on Anthropic's allegations.
Yahoo FinanceAnthropic Accuses Alibaba of Using 25,000 Fake Accounts to Copy Its AI; Stock Slides
Anthropic, an AI safety-focused lab, has accused Chinese tech giant Alibaba of executing what it calls 'the largest known distillation attack' on its Claude AI models. According to Anthropic, Alibaba used approximately 25,000 fraudulent accounts to harvest over 28.8 million exchanges with Claude between April and June 2026. The accusation was conveyed in letters to U.S. Senators Elizabeth Warren and Tim Scott, elevating the IP dispute into a potential national-security issue. The news caused Alibaba's ADR shares to drop 3.99% on the day, compounding a 38% year-to-date decline. The article frames the incident within the broader context of U.S.-China AI competition and notes that Alibaba's Qwen model family has over 1 billion downloads. Anthropic's Fable 5 Mythos models are currently suspended, partly due to this attack.
Yahoo FinanceAnthropic Accuses Alibaba of Massive AI Distillation Attack Using 25,000 Fake Accounts
Anthropic has accused Chinese e-commerce and AI giant Alibaba of orchestrating the 'largest known distillation attack' on its Claude AI models, involving roughly 25,000 fraudulent accounts that harvested over 28.8 million exchanges between April and June 2026. The allegations were detailed in letters sent to U.S. Senators Elizabeth Warren and Tim Scott, transforming the intellectual property dispute into a national-security matter amid heightened U.S.-China AI competition. Alibaba, which pitches itself as China's full-stack AI champion via its Qwen model family, has seen its stock decline 38% year to date, with shares falling another 4% on the day the news broke. Anthropic is effectively asking lawmakers to treat model distillation by a foreign adversary as a security issue, moving beyond typical commercial IP disputes. The incident also highlights the strategic importance of Alibaba's AI business, which reported 38% year-over-year Cloud Intelligence Group revenue growth in Q4 FY2026.
Yahoo FinanceAnthropic accuses Alibaba of using 25,000 fake accounts to distill its Claude AI model
Anthropic, a leading American AI lab, has accused Chinese tech giant Alibaba of illicitly using its Claude AI model to train its own models through a method called distillation. In a letter to the U.S. Senate Banking Committee, Anthropic claimed that over 25,000 fraudulent accounts were used in 28.8 million exchanges between April and June 2026, traced to operators linked to Alibaba and its AI lab Qwen. The accusation follows similar claims against other Chinese AI labs earlier this year. Anthropic warns that such practices could help China develop a frontier AI model rivaling its own Mythos Preview, a concern shared by U.S. lawmakers. The report highlights the intensifying U.S.-China AI race, with both sides using export controls and other measures to gain an advantage.
Latest from Tom's HardwareAnthropic Accuses Alibaba of Illicitly Distilling Claude AI Models
Anthropic, the American AI lab behind Claude, has accused Chinese tech giant Alibaba of using Claude's outputs to train its own AI models through a large-scale distillation campaign. In a letter to U.S. Senators Tim Scott and Elizabeth Warren, Anthropic claims the operation involved over 25,000 fake accounts and 28.8 million exchanges between April and June 2026, linked to Alibaba's Qwen AI lab. This follows similar accusations earlier in the year against DeepSeek, Moonshot, and MiniMax. Anthropic warns that such distillation could help China develop advanced AI models faster and cheaper, potentially matching Anthropic's own capabilities. The allegation comes amid intensifying U.S.-China competition for AI supremacy, with Washington imposing export controls on advanced chips and Beijing restricting rare earth materials.
Latest from Tom's HardwareAnthropic Accuses Alibaba of Largest Known AI Distillation Attack
Anthropic, the AI company behind Claude, has accused Chinese tech giant Alibaba of perpetrating the largest known distillation attack on its AI models. In a letter to US Senators Tim Scott and Elizabeth Warren, Anthropic's head of policy Sarah Heck stated that Alibaba-affiliated operators conducted 28.8 million exchanges with Claude through nearly 25,000 fraudulent accounts between April 22 and June 5. The operators allegedly sought to illicitly extract Claude's capabilities to train Alibaba's own AI models, including its Qwen LLMs. Anthropic called for legislation to limit China's access to US computing infrastructure and penalize Chinese entities launching such attacks. The accusation comes weeks after the US government imposed export controls on Anthropic's Fable 5 model. Alibaba, recently added to a Pentagon blacklist of Chinese military-linked companies, has sued the US government over the designation. Alibaba's share price dropped over 4% following the news.
All Content from Business InsiderAnthropic Accuses Alibaba of Illicitly Extracting Claude AI Model Capabilities
U.S. AI company Anthropic accused Chinese tech giant Alibaba of conducting a large-scale 'distillation' attack to illicitly extract capabilities from its Claude AI model between April 22 and June 5, 2026. Anthropic stated the operation involved over 28.8 million exchanges with Claude through nearly 25,000 fraudulent accounts, making it the largest known attack of its kind on the company. The accusation was detailed in a June 10 letter to U.S. Senators Tim Scott and Elizabeth Warren. Anthropic said the distillation effort aimed to accelerate China's ability to reach its advanced Mythos Preview capabilities. The article notes that Alibaba was added to the Pentagon's Chinese military companies list that month, and that the U.S. Commerce Department later imposed controversial restrictions on Anthropic's own Mythos and Fable AI models, citing national security concerns.
The Hindu: Latest News today from India and the World, Breaking news, Top Headlines and Trending News Videos.Anthropic Accuses Alibaba of Illicitly Accessing AI Models
Anthropic, an AI startup, has publicly accused Alibaba of illicitly accessing its AI models. The accusation comes amid broader concerns from Western AI firms, including OpenAI, that Chinese AI startups are using a technique known as 'distillation' to replicate their models. Anthropic is calling for additional government support to counter what it views as unfair competitive practices by Chinese rivals. The report published by The Business Times on June 25, 2026, highlights ongoing tensions in the global AI industry over intellectual property and competitive espionage, with implications for trade and technology policy.
The Business TimesAnthropic Accuses Alibaba of Massive AI Capabilities Extraction Campaign
Anthropic, a leading AI company, sent a letter to U.S. Senate committee leaders accusing Chinese tech giant Alibaba of orchestrating 'the largest known distillation attack' on its AI models. The letter, obtained by CNBC, alleges that operators affiliated with Alibaba and its AI lab used roughly 25,000 fraudulent accounts to conduct 28.8 million exchanges with Anthropic's models between April 22 and June 5, 2026. Distillation is a method where a smaller AI model is trained using outputs from a larger one. Anthropic stated that Alibaba ignored warnings from the Trump administration. The accusation comes amid broader U.S.-China tech tensions, with the Trump administration previously issuing an export control directive forcing Anthropic to suspend access to its latest Claude models by foreign nationals. Alibaba did not immediately respond to a request for comment.
US Top News and AnalysisAnthropic Accuses Chinese AI Labs of Industrial-Scale Model Distillation Attacks
Anthropic, the maker of the Claude chatbot, has formally accused three Chinese AI labs—DeepSeek, Moonshot, and MiniMax—of conducting industrial-scale distillation attacks to steal intellectual property from its flagship model. The company claims the labs generated over 16 million exchanges with Claude using 24,000 fraudulent accounts routed through commercial proxy services, effectively bypassing rate limits and detection systems. Distillation involves querying a superior 'teacher' model millions of times to train a cheaper 'student' model without paying initial training costs. Anthropic is now shifting from network-level defenses to behavioral analysis to detect such attacks. The accusations highlight growing tensions over AI intellectual property theft and the circumvention of U.S. export controls on advanced semiconductors.
The Hindu: Latest News today from India and the World, Breaking news, Top Headlines and Trending News Videos.