In a recent appearance on the PBD Podcast, Jacob Coxon, who previously held research roles at both Anthropic and OpenAI, stated he is 90 percent certain that China has successfully placed spies within top-tier US artificial intelligence companies. Coxon argued that such infiltration would allow China to bypass the time-consuming research and development phases, effectively closing the technological gap between the two nations by stealing proprietary advancements.

Coxon’s claims center on the strategic importance of AI as a national security asset. He suggested that if US labs were to pause their development, China would still be able to achieve parity through the acquisition of stolen intellectual property. However, the report notes that Coxon did not present any concrete evidence or documentation to substantiate these allegations during the interview. Neither Anthropic nor OpenAI has issued a formal response to these specific claims as of this writing. The assertion highlights growing concerns among some industry insiders regarding the security of sensitive AI research, though it remains a speculative claim from a single former employee.