AI
Chinese military researchers distilled outputs from US AI models

A a report review of more than 80 Chinese papers and patents found military-linked researchers using outputs from models including GPT-3.5 and Claude 3 Haiku to train smaller specialised systems. Reported applications included local code analysis, content monitoring and edge image processing. Distillation transfers selected behaviours; it does not reproduce a frontier model or remove the need for substantial compute.
Strategic impact: Model providers need stronger abuse detection and contractual controls, while governments face difficult distinctions between ordinary research, unauthorised extraction and export-control circumvention.




