A recent demonstration by cybersecurity researchers in California has highlighted the potential risks artificial intelligence poses to critical digital infrastructure, drawing attention from both Chinese and American experts. The researchers developed a tool named WeWorm, which could simulate a large-scale breach of WeChat, China’s ubiquitous messaging app with over 1.4 billion monthly users. The tool was capable of hijacking a user’s account, accessing contacts, and spreading from phone to phone without user interaction. The vulnerability was disclosed to Tencent, WeChat’s parent company, and U.S. authorities, with Tencent reportedly addressing the issue promptly.

WeChat, deeply integrated into daily communication, government services, and digital payments in China, is considered vital to the country’s digital ecosystem. The demonstration underscored how AI-powered cyberattacks could rapidly undermine platforms central to millions. Experts warn that AI’s increasing capabilities are outpacing traditional cybersecurity defenses, raising concerns that relatively small, non-governmental groups could mount highly damaging attacks. Zhao Minghao, deputy director of the Center for American Studies at Fudan University, described such an attack as a new kind of “nuclear weapon” in the digital age.

The issue of AI-driven cyber threats is expected to be part of discussions between President Joe Biden and China’s President Xi Jinping during their meeting scheduled for September 24 in Washington. While both countries acknowledge the growing risks, mutual suspicion remains a significant obstacle to collaboration. While China emphasizes the need for an open, consensus-based framework that involves multiple countries rather than dominance by the United States, the U.S. is advocating for greater transparency and risk mitigation measures.

Chinese officials have voiced concerns about AI’s use beyond hacking, emphasizing risks related to social stability. Minister of State Security Chen Yixin highlighted threats from AI-generated deepfakes, misinformation campaigns, and cognitive warfare, which could disrupt societal cohesion. Chinese regulatory bodies have issued guidelines to oversee AI’s development and use, with President Xi Jinping stressing the importance of international cooperation to ensure AI remains under human control.

Meanwhile, tech companies have also voiced alarm about AI-related risks. Anthropic, a prominent AI research firm, reported attempts by various actors, including a Chinese arms manufacturer, to misuse its AI models for military and surveillance purposes. These developments have fueled calls from industry leaders such as Bill Gates for global prioritization of AI risk management.

Despite ongoing tensions, the WeWorm incident has demonstrated that cross-border collaboration is possible and can be effective in mitigating cyber threats. The cooperation between the Californian researchers and Tencent in quickly addressing the vulnerability is seen by some experts as a positive example of how information sharing may help reduce the growing security risks associated with AI technologies.

Nonetheless, skepticism remains about whether substantive limits on AI development will be agreed upon. Analysts suggest that both China and the United States are unlikely to restrict their AI advancements significantly, given the strategic importance attached to the technology. Instead, establishing communication channels and frameworks to manage crises and share information about emerging threats are viewed as more realistic steps toward reducing the potential for conflict or major cyber disruptions fueled by AI advancements.