Geedge Networks, the Chinese firm behind much of the "Great Firewall," is developing AI systems that try to predict who could become a critic of the state before they act, according to a New York Times report based on leaked corporate documents analyzed by Vanderbilt University. The tools combine location data, telecommunications records and internet activity to build behavioral profiles and flag citizens deemed a political risk.
The findings rest on roughly 100,000 internal files, including technical specs, source code and client contracts, that leaked to a consortium of journalists and researchers. Geedge and its government-linked MESA Lab have already exported network-surveillance software to countries such as Ethiopia, Kazakhstan, Myanmar and Pakistan.
U.S. officials said there is no evidence the predictive system has been deployed yet, and noted that American chip export limits may have slowed development. Still, analysts warn the project signals a shift from monitoring what people say to preemptively identifying who might dissent.