China's delegation urges cooperation on AI governance at Munich meeting

Who is winning the global artificial intelligence race? This question draws active discussions at the 62nd Munich Security Conference, where policymakers, scientists and technology leaders gather for the event's first public forum dedicated to AI.

Fu Ying, former Chinese vice-foreign minister, led a delegation of Chinese scholars to the conference and spoke at the main session on Feb 13 titled "Outsmarting Ourselves? Risks and Rewards of the AI Race".

She outlined China's approach to AI governance and cautioned against framing technological development as a zero-sum contest.

China's basic principle in governing AI is to find a balance between innovation and safety, Fu said. Over the past decade, China has established a comprehensive legal framework for internet governance, covering cybersecurity, data security and personal information protection.

Based on this experience, China has integrated AI regulations into laws, giving emphasis to ethical standards and safety oversight for AI while supporting algorithm research and computing infrastructure.

Fu described China's regulatory approach as "agile governance", a gradual and adaptive method that introduces targeted measures in step with technological progress. Such an approach, she said, ensures that regulation keeps pace with innovation without stifling it.

Addressing the notion of "winning" the AI race, Fu said the term implies a finish line, whereas AI innovation is a continuing and evolving process.

"In science and technology, especially in AI, the latest model is always the best, and the next one will be better," she said. No single breakthrough can guarantee permanent leadership.

Chinese and United States companies have their own strengths in AI development, Fu said. The ultimate purpose of technological progress is to improve people's lives and enhance social well-being, she said. If major countries and leading companies fail to coordinate and cooperate, not only could technological progress suffer, but the risks associated with rapid development could increase if safety is sacrificed for speed.

She recalled a 2019 discussion with former US secretary of state Henry Kissinger at the New Economy Forum in Beijing, where he had expressed concern about the potential existential threats posed by disruptive technologies.

They agreed that if China and the US can cooperate, they may find ways to manage such risks together. Conversely, if countries pursue confrontational competition, "the machines may win", they said.

On international cooperation, Fu emphasized China's support for the United Nations in building global consensus on AI safety. She called for greater investment in capacity building and multilateral efforts to narrow the digital divide, as well as coordinated approaches to key issues of AI development, security and governance.

Other panelists echoed the urgency of the issue. Yoshua Bengio, a Turing Award laureate and director of the Montreal Institute for Learning Algorithms, warned that humanity is building increasingly intelligent machines while safety lags far behind.

Competitive pressures create a "prisoner's dilemma" in which actors may cut corners on safety to avoid falling behind, he said. "Without international coordination, all sides could lose."

Henna Virkkunen, executive vice-president of the European Commission responsible for technological sovereignty, security and democracy, said AI is "a critical technology" and that Europe must develop its own capabilities rather than depend entirely on other countries or a single company.

The true winners "will not necessarily be those who dominate AI, but those who can generate productivity gains while ensuring safety and earning public trust", she said.

She also underscored the importance of inclusive governance to prevent AI from being monopolized and to ensure developing countries actively shape its future.

US Senator Christopher Murphy, a member of the Senate Foreign Relations Committee, argued that winning the AI race should not be defined by capturing the best, but by avoiding the worst outcomes.

The US faces challenges as powerful lobbying by the AI industry has complicated domestic regulatory discussions, making both national and international governance efforts more difficult, he said.

AI was a major topic throughout this year's Munich Security Conference, with additional sessions exploring strategic algorithms and the prospect of an AI arms race.

Fu concluded that China remains committed to globalization and multilateralism, advocating an open and cooperative technological ecosystem. The future global technology landscape, she suggested, is more likely to be multipolar and highly interconnected, rather than divided into rival blocs.