Florida Republican Congresswoman Kat Cammack claimed at The Hill’s Nation Summit that the Chinese government has deployed AI-generated “digital twins,” essentially data-driven behavioral replicas, of members of Congress built from publicly available information. Dan Proft noted that Cammack did not provide specific supporting evidence for the claim, while framing the concept as a technological evolution beyond more traditional espionage methods involving human intelligence assets.
Proft welcomed Maximilian Garley, who works on special projects at the advisory firm Quint Strategy and authored a recent Spectator piece examining the concept, titled “Has China Cloned Your Congressman?” Garley explained that unlike traditional intelligence-gathering methods relying on direct personal relationships or compromised sources, digital twin construction relies on scraping publicly available material, including speeches, votes, social media posts, and media appearances, to build detailed behavioral and decision-making profiles of lawmakers. He said such profiles could theoretically be used to simulate and predict how officials might respond to specific policy scenarios or negotiating situations, offering foreign intelligence services a low-cost method for anticipating American political behavior.
Garley noted that such efforts would likely extend well beyond China-specific policy questions, since more comprehensive behavioral data would improve predictive accuracy across a wide range of related foreign policy areas. He said the most significant potential application lies in the intelligence and espionage space, where foreign operatives could theoretically rehearse recruitment or approach strategies against a simulated version of a target repeatedly before attempting real-world contact.
Asked whether the United States is likely engaged in comparable efforts, Garley said it remains unclear publicly, but suggested that if China is pursuing such capabilities, it is reasonable to assume American intelligence agencies are pursuing similar or more advanced techniques. Proft also raised Cammack’s related comments calling for stronger vetting standards for congressional staff, citing past incidents involving foreign intelligence infiltration of American political offices, including a previously reported case involving a Chinese Communist Party-linked staffer connected to New York City’s Adams administration. Gierlach agreed that traditional human intelligence assets remain highly valuable specifically because they can access private information unavailable through public data scraping alone, meaning the rise of AI-driven profiling does not eliminate the ongoing risk posed by human infiltration.
Garley closed by noting that the fundamental asymmetry in this dynamic stems from the openness of American democratic governance compared to the far greater secrecy surrounding Chinese political figures, meaning foreign adversaries likely have access to significantly more information about American officials than the reverse. He argued that reducing the risk by limiting public disclosure would run counter to core democratic values, concluding that greater transparency, while creating this vulnerability, remains a defining and worthwhile feature of the American system.