As artificial intelligence (AI) shifts from an "auxiliary tool" to an "intelligent partner" in daily life, youth in China are pressing for clear governance, highlighting that they will live longest with its consequences.
They stress that human agency must remain central, data barriers be broken, and intelligent agents properly constrained. These expectations set the tone ahead of the 2026 World AI Conference and High Level Meeting on Global AI Governance in Shanghai, scheduled from Friday through Monday.
Feng Erhu, assistant professor at the School of Artificial Intelligence at Shanghai Jiao Tong University, stressed that AI must remain under control, with human beings making the final decisions.
"AI cannot replace human beings in making final decisions. A controlled AI is a good AI. Of course, in this process we need to strike a better balance - determining what tasks AI can undertake and what boundaries AI can operate within. We need to build a more comprehensive governance system oriented toward large models and AI," he said.
On governance cooperation outcomes expected from the conference, Guo Muze, a 2024 master's student at the National School of Engineering Excellence at Tongji University, expressed hope for progress on data sharing and standardization.
"I am quite looking forward to seeing more concrete outcomes from this conference regarding data sharing and data norms. AI fundamentally requires the feeding of raw data. Our country leads internationally in infrastructure construction and can provide excellent data samples and references for global development, at least in the field of artificial intelligence and civil engineering, breaking down data silos," he said.
Xu Kehao, a student at the School of Computing and Intelligent Innovation at Fudan University, highlighted concerns about agent safety.
"What I am most concerned about is the safety of intelligent agents. In programming, we often use an agent to provide recommendations for various processes. Sometimes it executes some strange commands on the terminal, leading to various engineering mistakes and errors. At such moments, how do we constrain this agent? How much authority should we grant it? I hope this year's conference will feature relevant discussions and experience sharing," he said.
Themed "AI Partnership for a Brighter Future," the 2026 World AI Conference will feature more than 140 forums and bring together 1,400 guests from home and abroad. It will span six sections: Conferences and Forums, Exhibitions and Showcases, Awards and Competitions, Application Experiences, Innovation Incubation, and Talent Attraction.
Chinese students call for human‑centered AI governance, defined boundaries
