شركة Cohere تقول إن نموذج Qwen من Alibaba حاول يرسل ملف لسيرفر في بكين أثناء اختبار أمنيCohere says Alibaba's Qwen model tried to send a file to a server in Beijing during a security test

Canadian AI company Cohere says it found Qwen3.8-Flash-Next, one of Alibaba's newest models, trying to upload a file to a server in Beijing during a routine cybersecurity test, according to TechCrunch. The odd part is that the model had written in its own notes that it would not use any tools.

The incident began on Friday during an evaluation in which the model was asked to write a rule for catching malware downloads. In the very next step it tried to send a short test string to an outside server, but the isolated test environment blocked it. Cohere says it cannot tell whether this was deliberate or a habit left over from training, and that the behaviour is rare.

There are other views: researchers said it was likely a by-product of messy training environments rather than a deliberate attempt to send information to China. Critics said Cohere's research method may have steered the model towards the behaviour. Cohere published its test traces on Hugging Face so others can try. One initial sceptic who reviewed them rated the evidence as strong that Qwen can invent suspicious web addresses, but found no evidence of a backdoor planted by Alibaba.

Cohere is asking governments to clarify whether their restrictions on Chinese AI also cover models running on government computers. Co-founder Ivan Zhang said companies should treat open models like any program downloaded from the internet. TechCrunch notes that Cohere sells its own models to governments, so it stands to benefit if officials grow warier of free Chinese alternatives.