According to a report by The New York Times on June 1, the Chinese tech company Geedge Networks, through its government-backed research institute Mesa Lab, is developing an AI-driven political risk prediction tool. Unlike traditional surveillance that only tracks known dissidents, this system aims to build personal profiles for ordinary citizens by collecting and analyzing location data and online behavior. Using AI, it identifies potential targets who “might” criticize the government — marking them as risks before they take any public action. The information comes from internal Geedge Networks documents leaked and reviewed by researchers at Vanderbilt University. The system is still in development. Geedge Networks is a major supplier of the commercial export version of China’s firewall, providing surveillance and content filtering infrastructure to authoritarian governments. Brett Goldstein, director of the Vanderbilt University “Problematic Questions Laboratory,” commented: “This is what happens when mass surveillance meets artificial intelligence.”
The research team noted that Geedge Networks’ researchers “are not just recording behavior patterns but trying to predict what citizens will do next and with whom” — turning vast amounts of ordinary behavioral data into raw material for inferring individuals’ future political leanings. Notably, export restrictions on US AI chips imposed during the Biden administration are believed to have materially hindered Geedge Networks’ development process, with the report suggesting this has somewhat slowed the advancement of China’s next-generation surveillance technology. Meanwhile, a recent independent report by the Australian Strategic Policy Institute (ASPI) also documented the significant upgrading of AI tools used by the Chinese government over the past two years for “automated censorship, enhanced surveillance, and preemptive suppression of dissent.” These systems can now predict public protests and monitor the emotional state of detainees.