China

Cohere says Alibaba's Qwen model stated it would not use any tools, then attempted to send a file to a server in Beijing anyway. Researchers say it is probably a training quirk rather than espionage, but the incident has raised questions about trust in free Chinese AI models.

AI Safety • TechCrunch AI

CrowdStrike says China-linked hackers used an AI tool called ARTEX, built with a DeepSeek LLM and another model, to target more than 200 organizations in 11 days and compromise at least nine. The campaign hit governments and businesses worldwide, according to the report.

Security • Techmeme

China Telecom's XingChen lab has open-sourced Xing4.0-29B-Intent, a model that classifies intents and fills structured arguments in one JSON output. The release includes 2.75 million training samples and an inference SDK.

Models • Pandaily