As artificial intelligence continues to reshape industries and daily life at remarkable speed, its future will depend as much on responsible governance as on technological progress, said experts and business leaders.

"AI has become an irreversible trend of our time, profoundly changing the way we produce, live and govern," said Wu Dong, chief engineer of the Cyberspace Administration of China, at a parallel forum on frontier AI models during the 2025 World Internet Conference in Wuzhen, Zhejiang province.

Wu said China will continue to prioritize AI development in its 15th Five-Year Plan period (2026-30) by focusing on fundamental research and breakthroughs in key technologies, promoting industrial growth and expanding real-world application scenarios.

Innovation, he added, must advance hand-in-hand with responsibility, supported by sound governance, reliable data quality and professional expertise.

AI safety has become a central topic as technology moves from the virtual world into real-world applications. Lu Wei, vice-chairman of the Cyber Security Association of China, described AI security as the lifeline of development.

"As AI expands from the digital world into the physical world, its risks are becoming more complex. We face not only technical vulnerabilities, but also ethical, privacy and systemic risks."

Lu cited the AI Safety Governance Framework 2.0, China's latest policy upgrade, as a sign of a shift from reactive regulation to proactive governance. Released in September, the framework highlights transparency, data protection and a collaborative governance model that spans borders, fields and industries.

"High-level security does not constrain innovation — it safeguards it," Lu said. "Only by embedding safety into every layer of the AI ecosystem can we ensure sustainable, high-quality development."

Undoubtedly, AI is already reshaping productivity and creativity across sectors. A joint report by the Yinxiang Biji Research Institute and the WIC Think Tank found that integrating AI models into real-world applications has become a growing global consensus.

"In manufacturing, AI-driven optimization has improved efficiency by 12 percent; in pharmaceuticals, it has halved new drug development cycles; and in education, AI large-model interaction is redefining personalized learning," said Qiao Qian, director of Yinxiang Biji Research Institute and vice-president of Yinxiang Biji.

As AI becomes more deeply embedded in critical systems, cybersecurity specialists warn of emerging vulnerabilities. Hu Zhenquan, president of Chinese cybersecurity company 360 Digital Security Group, said AI security represents the next frontier of the digital economy.

"Everything can be programmed, imitated, generated and orchestrated — and that's both AI's power and its weakness," Hu said. He proposed an "AI + Security "dual strategy that uses AI to safeguard AI systems.

Beyond technology, experts emphasized the importance of shared governance and global collaboration in shaping a sustainable AI future.

"Digital transformation is irreversible — like biological evolution," said Gerald Vernez, founder and director of the digiVolution Foundation and former Swiss cyber defense representative. "The real issue is not the tools themselves, but how we use them."

Vernez stressed that "strategy must precede technology" and urged nations to strengthen "digital sovereignty built on trust, resilience and cooperation".

As large models proliferate, infrastructure resilience is becoming a new priority. Hou Shengli, chief technology officer of Cisco Greater China, said full-lifecycle governance — from vulnerability detection to data integrity — should form the backbone of AI infrastructure.

"Future systems will evolve across three dimensions — within servers, within data centers and between data centers — all supported by secure, high-speed networks," Hou said.