appConfigEntityId

bodyType

customContents

entityId

entityUuid

isVideo

articlePageQuery

article

articleSchemaContent

content

hooksArticlePageAdvertisementConfigArticle

hooksArticlePageFloatingButtonConfigArticle

appConfig

filter

json

liveArticle

Wei Wei

An American company found itself under attack by American artificial intelligence systems. Its unlikely rescuer was a Chinese AI model.
It sounds like a Hollywood science fiction film. It’s not.
During an internal test, advanced OpenAI models bypassed network restrictions, gained access to the internet and launched autonomous cyberattacks on open-source platform Hugging Face. As engineers rushed to respond, several leading closed-source AI systems reportedly failed to provide meaningful help because their security rules could not distinguish the attacker from the victim.
Hugging Face ultimately relied on a locally deployed Chinese open-source model – Zhipu AI’s GLM-5.2 – to analyse the attack, trace abnormal behaviour and support the emergency response. The timing was ironic.
For months, Washington has portrayed Chinese open-source AI models as a security threat. US policymakers have even proposed restricting or sanctioning their global deployment. Yet in a real AI security breach, it was an open-source Chinese model – not a proprietary American one – that proved useful.
Whether this is a historical turning point remains to be seen. But it raises an important question. The greatest challenge in AI may no longer be who builds the most powerful models, but who builds the most effective system to govern them.




The AI attack exposed three weaknesses in governance today. The first is the assumption that closed systems are inherently safer. Much of the AI industry has treated open-source development as a security liability. But when critical systems remain closed, outside experts cannot inspect them, identify vulnerabilities or help respond when something goes wrong. Transparency is not the enemy of security. Under the right governance framework, it can become one of its strongest safeguards.
China has followed a different path. As its AI industry expanded, it encouraged the development of open-source ecosystems while introducing regulatory oversight covering model security, risk assessment and accountability. The combination is significant. Openness allows broader participation in improving safety. Governance keeps that openness from becoming disorder.
The second weakness is that AI capability has advanced much faster than AI governance. Competition has focused overwhelmingly on building larger, faster and more capable models. Safety mechanisms have struggled to keep pace. Many companies have developed sophisticated safeguards against external attacks while paying far less attention to the possibility that increasingly autonomous AI systems could themselves become a risk.
China anticipated these challenges earlier than many observers recognise. Through regulations on generative AI services, AI security frameworks and technical standards, policymakers have increasingly emphasised that innovation and safety must advance together. The objective is to ensure that increasingly capable AI systems remain subject to meaningful human oversight.


The third weakness is that AI risks do not stop at national borders. An autonomous AI system can interact with platforms, networks and users around the world in seconds. A cyber incident in one country can quickly become a global problem. Yet AI governance remains fragmented. Different countries are developing separate regulatory systems, incompatible technical standards and isolated emergency response mechanisms.
Technology has become global. Governance should be the same. Conversation must move beyond competition. The question facing the international community is whether we can build a governance framework capable of responding to risks that are increasingly transnational.
This is where China’s Global AI Governance Initiative deserves greater attention. Proposed in 2023 and advanced at the recent World AI Conference, the initiative rests on a simple principle: AI should remain a global public good rather than becoming another source of geopolitical division.
It advocates balancing innovation with security, promoting open cooperation instead of technological monopolies, strengthening international co-ordination on risk management, and ensuring AI always remains under human control.


More importantly, China has begun translating these principles into practice. Domestically, besides its regulations covering generative AI, model evaluation and ethical governance, it is also accelerating work on national AI safety standards.
Internationally, China has expanded support for open-source AI development, proposed greater cooperation under the United Nations framework and announced new training programmes and AI cooperation platforms for developing countries.
These efforts do not claim to offer a perfect solution. No country possesses one. But they address a problem that has become increasingly difficult to ignore: AI governance cannot remain the responsibility of a handful of companies or a small group of technologically advanced nations.
Had the recent AI attack spread across interconnected platforms, its consequences would not have been contained by national borders. Critical infrastructure, financial systems and public services worldwide are becoming increasingly dependent on AI.


Future incidents may involve autonomous agents operating across multiple jurisdictions simultaneously. Without common standards for information sharing, incident response and technical cooperation, governments will always be reacting after the damage.
AI is entering an era in which its opportunities and risks are equally global. Competition will continue. Innovation should as well. But neither can replace governance. When AI itself begins creating risks that no single company, or even country, can fully control, international cooperation becomes less an ideal than a necessity.
That an American company had to be rescued by Chinese AI in an AI security crisis is more than just irony. It is a reminder that AI needs no passport to travel. Neither do its risks. If the world is serious about ensuring that AI serves humanity rather than threatens it, geopolitical rivalry cannot be the final answer. A more open, coordinated and inclusive approach to global AI governance is indispensable.


B-2 Spirit, AI cooperation, no US-China conflict: key Xi-Trump takeaways so far

China tech faces fresh scrutiny as US blacklist bill introduced, Lenovo case emerges

US regulators back multi-cancer blood test – but Hong Kong doctors remain cautious

Were smart glasses used in Singapore to film 2 female students?

Technology

How a Chinese model stopped a cyberattack when US guard rails failed

OpenAI admitted its latest flagship models – including GPT-5.6 Sol and an unreleased, “even more capable” system – recently breached Hugging Face’s infrastructure after escaping its testing sandbox. Photo: Reuters

US-China AI rivalry becomes a battle to export competing tech governance visions

President Xi Jinping joins attendees for a group photo before the opening ceremony of the World AI Conference in China on Friday. Photo: Reuters

Are the US’ AI models better than China’s? That may be beside the point

People try products at AI+ Power 2026, Hong Kong’s largest artificial intelligence expo and conference, at the Hong Kong Convention and Exhibition Centre in Wan Chai on June 4. Photo: Jonathan Wong

Zhipu AI exhibiting at the 2024 World Artificial Intelligence Conference in Shanghai. The Chinese company is seen as a champion of open-weight models amid global frustrations over Anthropic’s restrictions. Photo: Zhipu

As US-Chinese AI model gap narrows, what next for Washington?

China’s latest AI model advancements have forced Washington to ponder its next move in the global artificial intelligence race. Image: Shutterstock

People walk past the Moonshot booth promoting its Kimi K3 AI model during the World Artificial Intelligence Conference in Shanghai on July 17. Photo: Reuters

More Americans think China is world leader in AI, new survey reveals

Findings from a Pew Research Centre survey reveal that more than one-third of Americans see China, rather than the United States, as the global leader in artificial intelligence. Photo: Shutterstock

Opinion

China Opinion

World Opinion

News

China Future Tech

China

World

Science & Technology

Closed models aren’t inherently safer. As AI capabilities outpace governance, nations must cooperate to manage growing risks


Closed models aren’t inherently safer. As AI capabilities outpace governance, nations must cooperate to manage growing risks.


Artificial intelligence (AI)

OpenAI

Z.ai

Cybersecurity

United States

AI: Companies

AI: Policy

AI: Geopolitics

US-China tech war

China technology

Expresses the views, ideas and conclusions of the author. (Eg. Enough of Crazy Rich Asians, think of crazy poor Hongkongers.)


How a Chinese model stopped a cyberattack when US guard rails failed

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}Closed models aren’t inherently safer. As AI capabilities outpace governance, nations must cooperate to manage growing risks

Closed models aren’t inherently safer. As AI capabilities outpace governance, nations must cooperate to manage growing risks