Published：Jul 2, 16:00Last updated：Jul 9, 09:31

Alibaba Bans Claude Code Over Security Risks

Jul. 2, 2026 16:00

Alibaba has banned the use of Claude Code internally after identifying security risks, including potential backdoors. The company has classified the software as high-risk and will enforce the ban starting July 10, recommending Qoder as an alternative. The decision follows broader concerns about AI security and investment trends, with Alibaba and Tencent reportedly eyeing investments in AI firm DeepSeek.

Summarized

8News

5Media

Jul. 8, 2026 16:00

Chinese AI models, priced significantly lower than U.S. counterparts, are drawing global investment. Reports suggest U.S. firms now use Chinese models for 30-46% of tasks, up from 4.5% last year. The shift is driven by cost savings and performance, challenging U.S. dominance in AI and prompting regulatory concerns.

Jul. 9, 2026 09:31

China's MIIT accused Anthropic’s Claude Code AI of having a backdoor that transmits user data without consent. The tool, used for coding assistance, allegedly shares location and identity details with remote servers. China warned of privacy, security, and intellectual property risks. Users were advised to uninstall vulnerable versions or upgrade. Anthropic, linked to the Pentagon, has not responded. China has restricted Claude's use due to security concerns.

Alibaba has announced a company-wide ban on Anthropic's Claude Code software, citing serious security vulnerabilities including hidden backdoors and unauthorized data transmission. The ban takes effect July 10, with employees required to uninstall all related products and transition to Alibaba's internal alternative, Qoder. The decision follows warnings from China's Ministry of Industry and Information Technology (MIIT) about the tool's potential risks to corporate data security and user privacy.China.org+2

🔍 Security Concerns Identified

MIIT revealed Claude Code contained a backdoor that could transmit user location and identity data to remote servers without consent. The software also allegedly included a hidden detection mechanism targeting Chinese users. Anthropic described these features as experimental and claims they've been removed, but Chinese authorities maintain concerns about privacy and intellectual property risks.Reuters+1

💼 Corporate Response

Alibaba classified Claude Code as high-risk software after a comprehensive evaluation, with the ban applying to all internal systems. The company has recommended Qoder as its preferred alternative coding assistant. Meanwhile, reports indicate Alibaba and Tencent are exploring investments in domestic AI firm DeepSeek as part of broader AI security strategies.China.org+1

🌐 Global AI Market Shifts

The dispute coincides with growing adoption of Chinese AI models internationally. US companies like Coinbase and Uber have increased usage of Chinese-developed tools such as GLM and DeepSeek, attracted by lower costs compared to Western alternatives. Reports indicate Chinese models now handle 30-46% of AI tasks for some US firms, up from 4.5% previously.Nikkei Asia+1

⚖️ Ongoing Tensions

The situation highlights escalating US-China tensions in AI development. Anthropic has accused Alibaba of operating fake accounts to train its models, while China restricts Claude's use citing security concerns. The conflict underscores competing visions for AI governance and technological sovereignty between the two nations.France 24+1

MIITQoderAlibabaDeepSeekAnthropic