Chinese AI companies defend their safety practices amid Western criticism, emphasizing cultural context and internal evaluation measures.

Ahead of major model releases around Lunar New Year, Chinese firms argue that their approaches to AI safety should not be judged solely through a Western lens.

In 2025, Chinese models narrowed the performance gap with US rivals, prompting calls for increased focus on AI risks such as misuse and misalignment.

Some industry insiders note that Chinese companies conduct safety evaluations internally, despite limited public disclosure.

Zhipu AI, a Beijing-based firm, has yet to publish a safety framework despite signing voluntary global AI safety commitments in 2024, citing cultural differences in values.

Other companies, like Shanghai-based MiniMax, have participated in international AI governance dialogues but have not released formal safety protocols.

Experts suggest that Chinese firms are subject to stricter regulations domestically, including mandatory content labeling, which may influence transparency.

🔗 Source: South China Morning Post

Recent Zhipu AI developments