The new artificial intelligence dialogue between the United States and China may appear distant from Tibet, but it raises an important question for Tibetans: who will teach future AI systems about Tibet?
Following the September 2026 meeting between President Donald Trump and Chinese President Xi Jinping, the two countries established a dialogue on advanced artificial intelligence, which they agreed to call “super intelligence.” The discussions are currently focused on the risks and benefits of emerging AI technology rather than establishing a common information system.
Yet the larger issue is already visible. Millions of people increasingly use AI instead of conventional search engines to research history, religion and politics. Someone wanting to learn about Tibet may simply ask an AI system whether Tibet was historically independent, what happened in 1959, who the Dalai Lama is, or why Tibetans have protested Chinese rule.
The answer depends heavily on the information available to the system and the rules governing how that information is interpreted.
This creates a serious challenge for Tibetans because Beijing produces enormous quantities of digital material through government websites, state media, universities, research institutions and official publications. Even terminology can shape the information environment. Chinese authorities increasingly use “Xizang” in English language government material instead of the internationally established term “Tibet.”
The danger is not merely that an AI system may provide inaccurate information. It may present politically contested terminology as if it were neutral.
Words such as “liberation,” “separatism,” “Xizang,” “Tibet,” “national unity” and “independence” carry very different historical and political meanings depending on the source.
China’s own rules for generative AI demonstrate how directly politics can enter this process. Chinese regulations require public AI services to uphold socialist core values and prohibit content considered to promote separatism or undermine national unity. These requirements apply not only to generated answers but also influence areas such as model development and training data selection.
This makes the question of Tibetan data particularly important.AI systems learn from what has been digitized. If Chinese government documents exist in enormous searchable databases while Tibetan historical records, exile archives, oral histories, Buddhist texts and Tibetan language scholarship remain poorly digitized, the imbalance could eventually affect what AI systems consider authoritative.
In the past, political narratives were shaped through newspapers, textbooks, television and universities.The machine itself becomes an intermediary between history and the person trying to understand it.
For Tibetan institutions, this means preserving history is no longer enough. Tibetan sources must also become accessible to the digital systems that will increasingly explain that history to the world.Archives need digitization. Tibetan language datasets need expansion. Historical documents need reliable translations. Tibetan scholarship and primary sources need to exist in formats that modern AI systems can discover and process.
The emerging US China AI dialogue is therefore worth watching not because Washington and Beijing have agreed on how Tibet should be represented. They have not.The deeper question is what happens as governments begin negotiating the rules of increasingly powerful artificial intelligence systems.
For Tibetans, the struggle over historical memory may soon extend beyond books, archives and classrooms into the databases and algorithms that determine what billions of people are told when they simply ask:
“What is Tibet?”
