Alibaba Is Said to Ban Staff From Using Claude Code After Alleged Security Risks Found(Yicai) July 6 -- Alibaba Group Holding is set to prohibit employees from using Claude Code in the workplace after alleged security risks -- specifically a potential backdoor vulnerability -- were reportedly identified in the artificial intelligence coding assistant developed by US-based Anthropic, according to an insider at the Chinese technology giant.
Alibaba has classified Claude Code as high-risk software after a comprehensive assessment and will prohibit its use from July 10, the source told Yicai on June 3. Employees are being recommended to switch to the Hangzhou-based company’s own agentic coding platform Qoder as an alternative, the person added.
Since its launch in May last year, Claude Code has become one of the world's most popular AI coding tools, with its revenue exceeding USD2.5 billion as of this February, according to Anthropic.
Anthropic has recently drawn criticism for blocking Chinese users. Since late last month, many have reported that their Claude accounts have been suspended without warning. Some developers have subsequently alleged that Anthropic has embedded a covert user-identification system within Claude Code to determine whether users are linked to China.
Anthropic also submitted materials to US officials last month, accusing Alibaba of conducting large-scale distillation attacks on its Claude model. The San Francisco-based company had earlier levelled similar accusations against Chinese AI startups DeepSeek, Moonshot AI, and MiniMax.
Alibaba previously encouraged staff to use both domestic and foreign AI tools, with programmers frequently using Claude Code, senior AI expert Guo Tao said to Yicai. But following Anthropic's accusation of model distillation attack, along with its targeted identification and suspension of Chinese users, Alibaba concluded that its research and development data could face leakage and traceability risks, so decided to ban the use of Claude Code, Guo noted.
The move by Alibaba could mark the start of a broader wave of risk-control measures regarding overseas AI tools among Chinese firms, with China’s internet sector likely to accelerate the shift to domestic coding alternatives, Guo stressed. This includes the development of in-house tools such as Alibaba’s Qoder, which is expected to see opportunities for large-scale deployment, Guo said.
Launched last August, Qoder had over five million users as of May, according to Alibaba. The platform can autonomously complete the entire software development workflow, from requirements analysis through code deployment.
Alibaba has been stepping up investment in AI since the end of last year, with Chief Executive Eddie Wu aiming to grow annual revenue from cloud computing and AI to USD100 billion from CNY100 billion (USD14.7 billion) within five years. The firm recently formed a Token Hub business group and a Token Foundry business unit to strengthen its AI organizational structure.
Editors: Dou Shicong, Martin Kadiev
