“China's long-term strategic planning and extensive corporate ecosystem have made it the largest partner of ITU-T SG17, the International Telecommunication Union's security standardization study group, while European stakeholders are seeking to deepen their engagement,” said Arnaud Taddi, Chair of ITU-T SG17, on the sidelines of the World Internet Conference Wuzhen Summit held in Zhejiang Province in eastern China during Nov 2-5.

“The vehicle is the TITA (Trust in Humans and Digital Identity) focus group, which tackles edge-computing security and digital identity interoperability jointly with Alibaba and other firms, in three steps: launch, joint research, and adoption. The intended output is a cross-domain, cross-platform interoperable digital identity standard,” he added.

“Chinese companies, with a solid economic and technological base and keeping delivering solutions in AI and IoT, provides a foundation that underpins our influence in international standard-setting.” he said.

Alongside standardization, AI safety has emerged as another key area in advancing bilateral cooperation.

Max Stauffer, co-founder and CEO of the Simon Institute for Longterm Governance, a Swiss think tank, pointed to the growing threat posed by AI-enabled cyberattacks and autonomous agent systems. With incidents such as coordinated attacks by groups of AI agents targeting websites becoming increasingly visible, cross-border cybersecurity risks are emerging as a shared concern.

China and Europe should therefore strengthen the exchange of information and data on such incidents, he argued, to better identify, assess and mitigate risks that affect both sides.

According to a report on risk perception and collaborative governance for AI safety and governance co-authored by AI Safety Connect and its Chinese and international partners and published during the conference, media reports on AI risks have grown 6 folds in the past six years.

Cyrus Hodes, a member of the OECD Expert Group on AI Futures and co-founder of AI Safety Connect proposes three steps, which are also the focus of talks with Chinese colleagues: prevent jailbreaks during development, conduct pre-deployment evaluation through an independent global assessment system, and build trust across ecosystems via computer navigation and similar techniques covering evaluation-to-verification.

"Countries around the world have agreed to establish a mechanism for reporting adverse events in AI laboratories. "We're encouraging work on open-source evaluation, incident reporting, critical-infrastructure protection and global compute moniotring to safely advance open source models in China, the US and beyond. The bottom line is to prevent critical infrastructure such as energy grids and hospitals being taken over by AI or being exploited by bad actors, especially when deploying in developing countries,” he said.

“The key focus of discussions with our Chinese colleagues is to build trust across ecosystems throughout the entire process from assessment to verification through technical means such as computer navigation. We expect to invite more Chinese colleagues to the Geneva summit in June to jointly discuss important AI safety issues," he added.

Legal cooperation is another important dimension of China-Europe engagement.

Prof. Bernd Holznagel of the University of Münster called for closer legal cooperation with China, noting that the rapid evolution of AI is creating regulatory and accountability challenges that no single country can address alone.

"AI models that once took a year to iterate can now be updated in a matter of weeks. Meanwhile, AI agents are displaying increasingly autonomous behavior, making them difficult to control completely. When an AI system makes a mistake, determining who should bear legal responsibility becomes a fundamental challenge," he said.

Cybersecurity and AI agent-related challenges transcend national borders and must therefore be addressed through international cooperation. A key task is to develop appropriate legal frameworks for frontier models and AI applications across different sectors," Holznagel noted.

Sectorwise, John Higgins, Chair of International AI Governance Association, said that initial efforts should focus on selected vertical industries, particularly the automotive and telecommunications sectors.

In this regard, China’s open-source AI has strategic implications for Europe, according to Prof. Dame Wendy Hall, a Fellow of the Royal Academy of Engineering and the Royal Society.

“The UK and many other European countries currently lack domestically developed foundational models at the frontier of AI and remain heavily reliant on US technologies. Against this backdrop, China's rapid innovation in open-weight models offers an important alternative,” she said.

Access to Chinese open-weight models could also give European countries greater technological flexibility and help reduce their exposure to potential political restrictions or technology-related barriers imposed by the United States, Hall noted.