ai
China's MIIT Starts Building AI Safety Benchmark for Generative Models
China’s Ministry of Industry and Information Technology has begun developing a safety benchmark to evaluate artificial intelligence models. The effort recruits companies and experts with applications due Tuesday after a notice issued Monday.
Jul 13, 9:30 AM·1m read1 source
South China Morning PostAudio version
Tap play to generate a narrated version.
Developing·Limited corroboration so far. This page will refresh as more sources emerge.
China’s Ministry of Industry and Information Technology has started building a safety benchmark to evaluate artificial intelligence models, South China Morning Post reported. The MIIT-led National Industrial Information Security Development Research Centre is recruiting companies and experts to co-build the benchmark. Applications are due on Tuesday after a notice published on Monday.
The institute stated that current frameworks fail to meet complex safety-governance needs and that a standardised testing platform is required to support industrial compliance. The benchmark will evaluate generative AI across six core dimensions: content safety, value alignment, robustness, fairness, privacy protection and trustworthiness.
A hybrid methodology will cover 31 specific safety risks across five major categories.
The system will combine automated fuzzing and stress testing with human oversight to control hallucination rates and data leaks and to curb jailbreak attacks on large language models.
