Open    Access



AI Translation Agent Empowers Domestic Game Localization From the Perspective of "Human-AI Collaboration"—A Case Study of Zhipu Qingyan Platform



Author(s)

GAO Yue, CHE Huan-huan

Full-Text PDF XML     21 Views

DOI:10.17265/1539-8080/2026.01.005

Leshan Normal University, Leshan, China

Against the backdrop of domestic games “going global” becoming a core path for the international communication of Chinese culture, the quality and efficiency of game localization directly affect the effectiveness of overseas market expansion. Traditional translation models suffer from high costs, long cycles, and unstable quality, while general artificial intelligence (AI) translation faces shortcomings, such as inconsistent terminology and poor cultural adaptation. Based on the concept of “human-AI collaboration”, this paper constructs an AI translation agent adapted to game localization scenarios using the Zhipu Qingyan platform. Through the construction of an exclusive knowledge base, customized workflow arrangement, and feedback optimization mechanism, it achieves dual improvements in translation efficiency and quality. Tests show that the agent increases translation efficiency by over 65%, the manual evaluation accuracy of cultural imagery transmission reaches 82%, the terminology consistency rate exceeds 92%, and the translation accuracy rate is 89%. It can shorten the translation cycle by 70% and reduce costs by more than 80%, providing an efficient and feasible technical solution for domestic game localization with significant practical value.

human-AI collaboration, AI translation agent, domestic game localization, Zhipu Qingyan

GAO Yue & CHE Huan-huan, AI Translation Agent Empowers Domestic Game Localization From the Perspective of “Human-AI Collaboration”—A Case Study of Zhipu Qingyan Platform. US-China Foreign Language, January 2026, Vol. 24, No. 1, 34-38 doi:10.17265/1539-8080/2026.01.005