At the 2026 World Artificial Intelligence Conference (WAIC) and the High-Level Meeting on Global AI Governance held in Shanghai from July 17 to 20, a forum focused on AI application innovation was hosted by TENCENT.

TENCENT Vice President Lin Songtao shared his perspective, stating that the gap in core AI model capabilities is narrowing, but the disparity in engineering implementation is widening. He noted that while differences in the reasoning performance of mainstream models are becoming more pronounced, the critical challenge lies in translating these capabilities into tangible business productivity and enhancing efficiency for individuals and organizations, which remains a key topic of discussion.

He elaborated that while models primarily define the upper limit of AI's potential, engineering determines its reliable baseline performance. Users may be initially attracted by a compelling feature, but their sustained usage and trust ultimately depend on the consistent, dependable quality of the underlying work.

Lin further expressed that a truly effective AI should not merely accelerate tasks but should also free up more human time for contemplation and creative endeavors. The vision is for AI to evolve into an inclusive productivity tool that is accessible to every individual and can be seamlessly integrated for collaboration within every organization, progressing from personal assistants to team and organizational-level solutions.