US and China move toward dialogue on artificial intelligence risks

Saturday 05 September 2026 - 12:00

By: Azzat Manal

US and China move toward dialogue on artificial intelligence risks

The United States and China are reportedly moving closer to establishing a new dialogue on the security risks associated with artificial intelligence, as the rapid development of advanced AI systems raises growing concerns about their potential use in sophisticated cyberattacks.

According to people familiar with the matter cited by Reuters, officials from the two countries are preparing for discussions in mid-September focused primarily on AI safety. If confirmed, the talks would represent a significant development in efforts to address the risks posed by increasingly capable AI systems and would mark the first such discussions between Washington and Beijing since Donald Trump returned to the White House for a second term.

US Treasury Secretary Scott Bessent is expected to lead the American side, although details regarding China’s representatives have not yet been disclosed. A White House official, however, said that no AI-related meeting had been formally scheduled for the middle of September, meaning the proposed timing could still change.

The initiative comes as cybersecurity experts increasingly warn that advanced AI tools could perform complex digital operations with limited human intervention. Washington has therefore explored possible forms of cooperation with Beijing aimed at identifying emerging threats and improving information-sharing mechanisms.

One proposal reportedly involves encouraging AI laboratories in both countries to monitor the systems they develop and exchange information about cyberattacks in which advanced AI models may be involved. Such cooperation would seek to improve early detection of malicious activity while creating communication channels between the two technological powers.

Recent incidents have highlighted the growing complexity of these risks. OpenAI and cybersecurity researchers have reported cases involving autonomous AI agents being used in attacks, raising concerns about how quickly such systems could evolve from tools assisting human operators into systems capable of carrying out substantial parts of cyber operations themselves.

Other reported incidents have further intensified concerns about the use of AI agents to compromise digital infrastructure and interact with other automated systems. These developments have added urgency to international discussions about safety standards, oversight and the limits that should apply to increasingly autonomous AI technologies.

The proposed US-China dialogue comes at a particularly sensitive moment, as artificial intelligence has become a major area of strategic competition between Washington and Beijing. Both countries are investing heavily in advanced models, computing infrastructure and AI-related research, while simultaneously seeking to protect their technological advantages.

US officials have also raised concerns about Chinese companies potentially using outputs from leading American AI models to accelerate the development of their own systems. One technique involved in such discussions is model distillation, a process in which the outputs of a larger and more sophisticated model can be used to help train a smaller system, potentially reducing the resources required to develop new AI capabilities.

Despite their rivalry, Washington and Beijing have strong incentives to maintain communication over risks that could extend beyond national borders. Cybersecurity threats, autonomous systems and the potential misuse of advanced AI models are challenges that could affect governments, companies and critical infrastructure worldwide.

The timing of the proposed discussions is also significant because Trump is expected to meet Chinese President Xi Jinping in Washington on September 24. The presidential meeting could provide an additional opportunity for AI safety to become part of the broader US-China agenda, alongside trade, technology and security issues.

Any formal dialogue on AI safety would not eliminate the strategic competition between the two countries. However, establishing channels for communication could help both sides better understand emerging threats and reduce the risk of misunderstandings as AI capabilities continue to advance.

For Washington and Beijing, the challenge will be to balance technological competition with the need to manage risks that neither country can address entirely on its own. A structured conversation on AI safety could therefore represent an initial step toward developing mechanisms for managing one of the most consequential technologies shaping the global economy and security landscape.