China has used an international workshop in Hangzhou to frame artificial intelligence as a potential public good for chemical safety and for implementation of the Chemical Weapons Convention. People’s Daily, republishing a September 2 Xinhua report, said Xie Yuansheng, vice minister at the Ministry of Industry and Information Technology, announced the proposal at the opening session of a four-day chemical-safety and security AI workshop.
Xie said China wanted to cooperate internationally on public AI tools that could support CWC-related chemical-weapons prevention, widen technology collaboration, and help developing countries narrow digital and intelligence gaps. The workshop’s theme was “Trustworthy AI for responsible use.” Its program listed 15 specialists from 15 Chinese and international organizations for online or in-person lectures, with nearly 50 attendees representing 31 countries.
The statement is not the announcement of a new binding treaty, funded international program, or deployed chemical-safety model. It is a ministerial proposal at a workshop. Its significance lies in the policy language China is using: AI is being presented as a shared capability for a sensitive international-governance issue, while the event emphasizes trustworthiness and responsible use.
The Workshop Links AI to Chemical-Safety Governance
The event began in Hangzhou on September 1 and was scheduled to run for four days. People’s Daily said it included keynote lectures, roundtables, tabletop exercises, and site visits. That format suggests a technical and policy exchange rather than a negotiating conference. The source does not identify specific AI systems, datasets, participating companies, or new products introduced at the workshop.
Xie’s remarks connected AI with two areas: the development of China’s chemical industry and implementation of the Chemical Weapons Convention, or CWC. The CWC took effect in 1997 and bans activities involving chemical weapons, including their production, possession, use, and transfer. AI does not alter those prohibitions. The policy question is how analytical tools, data systems, and related technologies might support safety, security, monitoring, or risk management in ways consistent with the convention.
The phrase “AI public goods” carries a specific diplomatic implication. It points toward capabilities that can be shared or used across countries rather than retained only by one company or state. The report does not define what form those public goods will take, whether they would be software, training, technical standards, or services, or when they might be delivered. It would be inaccurate to treat the phrase as proof that a particular open-source project or international platform already exists.
China has used similar language in other AI diplomacy settings, including its recent proposal for an SCO AI applications center. EastFrontier covered that initiative in its article on China’s AI diplomacy at the SCO summit. The chemical-safety workshop is more specialized. Instead of speaking about AI cooperation in general terms, it ties the concept to a treaty-based framework and to the management of chemical safety and security risks.
Trustworthy AI Is the Event’s Stated Organizing Principle
The workshop’s theme, “Trustworthy AI for responsible use,” gives the statement a governance dimension. Chemical safety is a field in which inaccurate recommendations, weak controls, or improper access to sensitive information can create real risks. The source does not set out a technical definition of trustworthy AI or a formal standard that participants adopted. It does show that China is placing responsible use at the center of an international discussion on AI in this domain.
People’s Daily reported Xie’s description of the “AI Plus” initiative and a planned CWC-related AI ecosystem that he said should be technology-enabled, coordinated among stakeholders, transparent, and trustworthy. Those terms are policy aspirations. They should not be mistaken for evidence that such an ecosystem has already been built or independently evaluated.
The practical challenge is complex. A chemical-safety AI system might help organize technical information, identify anomalies, support training, or assist experts with large document collections. But any real use would require decisions about data quality, access, oversight, and the limits of automation. The workshop report does not disclose a particular use case or model, so it does not establish which of these applications China or its partners will pursue.
That caution is especially important because the proposal sits beside broader international debates about AI safety. EastFrontier’s analysis of the U.S.-China AI safety-talk window focused on a different question: whether major powers can create channels for discussion of frontier AI. The Hangzhou event is narrower and more operational. It concerns chemical safety and the CWC, not a bilateral arrangement on general model safety.
International Participation Gives the Proposal a Practical Test
The reported participation provides a sense of the workshop’s reach. Nearly 50 people from 31 countries attended, while 15 experts from 15 domestic and international organizations gave lectures either in person or online. These figures show an international setting, but they do not identify which countries endorsed China’s proposal or what commitments, if any, were made.
Tabletop exercises and site visits could help participants discuss how an AI-enabled tool might be used in a controlled setting. Yet the report does not disclose the scenarios, conclusions, or follow-up plan. There is no published commitment to fund an AI public good, no technical specification, and no timetable for a shared platform. The responsible conclusion is that China has placed the proposal on the agenda, not that it has completed it.
The event nevertheless reveals a wider direction in Chinese AI policy. Officials are pairing the promotion of AI applications with language about safety, trust, international cooperation, and benefits for developing countries. That is a different emphasis from domestic industrial targets or model launches, but it can influence how Chinese agencies present AI in multilateral settings.
For chemical safety, the value of any AI public good will be determined by implementation. It would need credible technical design, clear governance, participation from international partners, and safeguards that respect the sensitivity of the CWC context. The September workshop provides the beginning of a policy conversation, with a specific theme and a stated willingness to cooperate.
China’s proposal is therefore best understood as a diplomatic and technical positioning move. It brings AI into a treaty-related conversation about chemical safety and security, but it does not yet create a new international mechanism. What follows from the workshop will show whether the phrase “AI public goods” becomes a defined program or remains an aspiration expressed at an opening session.