Alibaba banned employees from using Claude Code
Chinese tech giant Alibaba has banned employees from using Anthropic's Claude Code at work, citing concerns over the tool's AI features that allow it to identify users with ties to China.
Chinese tech giant Alibaba has banned employees from using Anthropic's Claude Code at work, citing concerns over the tool's AI features that allow it to identify users with ties to China.
This ban was part of an escalating conflict between the two companies after Anthropic accused Alibaba of illegally copying the capabilities of its artificial intelligence model Claude, Reuters writes .
Claude Code has become popular among programmers in China, despite Anthropic's restrictions on access for Chinese users and organizations.
According to a Reuters source, Alibaba employees were ordered to use the company's own coding platform called Qoder.
Last month, Anthropic said it had been attacked by Alibaba, which it described as an attempt at “distillation,” a process that involves training a weaker AI model on the responses (outputs) of a stronger one.
A letter to two US senators, seen by Reuters, says such distillation helps accelerate China's ability to achieve the advanced capabilities of Anthropic's Mythos Preview model.
Alibaba's ban came just days after developers discovered that Claude Code contained mechanisms that checked the user's environment (including time zone and proxy information) and inserted hidden tokens into prompts sent to Anthropic's servers.
On Tuesday, an Anthropic employee wrote on social media X that the feature was an “experiment launched in March” that was intended to prevent abuse by unauthorized account re-sellers and protect the model from distillation.
A Reuters source who spoke to Alibaba about the ban said Anthropic's anti-China restrictions are difficult to enforce against individual users because they can deploy servers in the United States, making their traffic appear to be coming from there.
However, companies are more aware of legal and regulatory risks, the source added.
While American AI model developers are struggling to prevent unauthorized access, resale, and distillation of their systems, Chinese cloud and AI companies have shifted their focus to in-house solutions and open-source models. These include DeepSeek, Alibaba's Qwen, Moonshot, and Zhipu.
At the same time, Chinese AI models are gradually penetrating the American market, a trend that has already raised concerns among some industry experts in the United States.
Alibaba integrates Qwen AI model into Taobao and launches agent shopping
Alibaba integrates Qwen AI model into Taobao and launches agent shopping
Alibaba introduces new Qwen3.5 model for the “Agent AI Era”
Alibaba introduces new Qwen3.5 model for the “Agent AI Era”
Ray-Ban from Meta has a new competitor: Alibaba releases Quark AI glasses
Ray-Ban from Meta has a new competitor: Alibaba releases Quark AI glasses
Read the country's main IT news in our Telegram
Read the country's main IT news in our Telegram
