A diplomatic rift over academic freedom is threatening to disrupt one of the world’s most influential gatherings of computer scientists. The China Association for Science and Technology (CAST) is initiating a boycott of the NeurIPS conference, a move that underscores a growing US–China divide in artificial intelligence research and highlights the increasing friction between geopolitical sanctions and scientific collaboration.
The conflict centers on a policy dispute regarding the eligibility of researchers from “US-sanctioned” institutions. While the organizers of NeurIPS—a US-based non-profit—eventually apologized and revised the rules, the fallout has persisted. CAST has signaled that it will no longer provide funding for researchers to attend the event and, more critically, will discount NeurIPS papers in the evaluations used to determine researcher standing and professional advancement in China.
This tension arrives at a pivotal moment for the global AI landscape. For the first time last year, China-based researchers accounted for the largest share of first authors at the NeurIPS conference held in San Diego. The scale of this influence is backed by data: a 2025 study by Digital Science indicates that China now publishes the highest volume of AI research papers globally, possessing a talent pool that significantly outpaces its competitors.
Controversy over attendance at the 2026 NeurIPS meeting reflects a growing divide in artificial-intelligence research between the two nations.Credit: Marciobnws/Shutterstock
The Timeline of the Policy Dispute
The friction began on March 23, when NeurIPS organizers announced a policy for the upcoming conference—scheduled for December 6 to 12 in Sydney, Australia—that would reject papers from researchers associated with institutions under US sanctions. To define these exclusions, the organizers provided a link to a US government webpage. This list included major Chinese technology firms, such as Huawei, which had previously served as a conference sponsor.
The announcement triggered an immediate backlash across social media and academic circles. Researchers from various nations expressed a willingness to boycott the event, arguing that scientific inquiry should remain independent of political sanctions. Two major Chinese bodies led the response: CAST, which manages funding for international academic exchanges, and the China Computer Federation (CCF), a primary professional organization for the nation’s computer scientists.
Following the outcry, NeurIPS issued an apology and amended the policy to apply to a significantly narrower list of entities. While the CCF welcomed this clarification, CAST remained firm. In a statement issued on March 31, CAST confirmed that its funding restrictions would remain in place, with resources originally earmarked for NeurIPS being redirected toward other domestic or international projects.
| Date | Event | Outcome |
|---|---|---|
| March 23 | NeurIPS announces sanctions policy | Papers from US-sanctioned entities face rejection |
| Late March | Academic and social media outcry | Calls for boycotts from CAST and CCF |
| Post-Outcry | NeurIPS issues apology | Policy narrowed to fewer institutions |
| March 31 | CAST maintains boycott | Funding redirected; papers discounted in evaluations |
Geopolitical Stakes and Scientific Autonomy
Analysts suggest that the boycott is less about a single conference policy and more about a strategic shift in how China views its role in the global AI ecosystem. Rebecca Arcesati, an AI policy expert at the Mercator Institute for China Studies in Brussels, suggests that CAST’s decision is a signal that China believes it can increasingly sustain its advanced AI research independently. According to Arcesati, the move demonstrates that if Chinese talent is not welcomed at traditional gatherings, the nation is prepared to forge its own paths.
The potential impact on the conference itself is substantial. William Hannas, a China specialist at Georgetown University’s Center for Security and Emerging Technology, noted that without Chinese participation, NeurIPS would be devastated. He suggested that if all Chinese co-authors were to withdraw their papers, the event would be gutted, given the sheer volume of contributions coming from Chinese institutions.
This sentiment is echoed by Yunji Chen, director of the State Key Laboratory of Processors in Beijing. Chen stated that the actions taken by CAST send a clear signal that “science should not be made a casualty of politics.” He argued that for NeurIPS to fully repair the damage, the organization must publicly reaffirm a commitment to academic neutrality.
What This Means for Global AI Research
The divide exposed by this boycott has implications that extend beyond the 2026 meeting in Sydney. For decades, the “open science” model—where researchers share breakthroughs regardless of nationality—has driven the rapid acceleration of deep learning and neural networks. The integration of US sanctions into conference admissions suggests a shift toward “siloed” research, where national security concerns override academic exchange.
For researchers, the stakes are personal and professional. In the Chinese academic system, publication in prestigious venues like NeurIPS is often a prerequisite for tenure, funding, and promotions. By discounting these papers in official evaluations, CAST is effectively lowering the professional value of the conference within China, potentially steering a generation of scientists toward domestic journals and alternative international forums.
The situation remains fluid, as it is not yet clear exactly how many individual researchers will follow the CAST directive or if the CCF’s more lenient stance will encourage some to attend regardless of state funding. However, the precedent sets a challenging tone for future collaborations in an era of heightened competition for AI supremacy.
The next critical checkpoint will be the paper submission and review cycle leading up to the December 6 start date in Sydney, which will reveal the actual extent of the withdrawal of Chinese research. Further updates on the status of academic neutrality commitments from the NeurIPS board are expected as the event approaches.
We invite readers to share their perspectives on the intersection of national security and scientific collaboration in the comments below.
