As President Trump and Chinese leader Xi Jinping prepare to meet in Washington next Thursday, experts acknowledge that the chances of resolving deep-seated mistrust over artificial intelligence are slim. The two nations remain locked in fierce competition, with the Trump administration recently accusing Chinese firms of stealing U.S. AI intellectual property through “distillation” and threatening sanctions. In response, Beijing has dismissed American chip export restrictions as a “Cold War playbook” designed to suppress Chinese technological advancement.
Despite this official hostility, a vibrant network of informal discussions known as “Track Two” diplomacy is actively shaping bilateral relations. These sub-national dialogues, often organized by think tanks and universities, bring together experts from both sides to discuss sensitive topics without the pressure of immediate political consensus.
Sarah Beran, incoming president of the National Committee on U.S.-China Relations, described AI as a critical yet challenging subject due to fears of economic chokepoints and weaponization. However, she noted that Track Two forums have historically served as effective backchannels. For instance, informal talks helped lay the groundwork for the restoration of military-to-military communications after they were severed following Nancy Pelosi’s 2022 visit to Taiwan, a move later formalized during the 2023 summit in Woodside, California.
Retired Air Force General Jack Shanahan highlighted that quiet Track Two engagements were instrumental in reaching the agreement between Biden and Xi that human control must be maintained over nuclear weapons decisions. Shanahan observed a shifting perspective among Chinese counterparts, who increasingly recognize that AI is too vital to ignore despite disparities in certain technological areas.
Samm Sacks of Johns Hopkins University, who participated in summer dialogues in Shanghai and Beijing, reported a strong sense of shared vulnerability regarding AI risks among participants. She emphasized that while governments argue over export controls and IP theft, experts on both sides acknowledge common ground on safety concerns such as cyber threats and bio-risk.
Scott Kennedy of the Center for Strategic and International Studies warned that these dialogues require government support to translate into policy. He described a current leadership vacuum, noting that while Track Two discussions can buy time and map out potential scenarios, they cannot independently resolve major geopolitical conflicts. Participants remain hopeful that signals from the upcoming Trump-Xi meeting will reinforce the importance of continued expert engagement on AI safety and stability.
Experts agree on risks but governments can’t cooperate? That leadership vacuum sounds dangerous. I wonder if these dialogues just delay inevitable conflict.
It is fascinating that military communications were saved by informal talks after Pelosi’s Taiwan visit. Maybe AI safety can follow that same path.