Somewhere between the speed of innovation and the slowness of institutional caution, a Chinese-built AI tool — allegedly constructed on stolen American intellectual property — found its way into a US government website, running quietly beneath the surface of public-facing infrastructure. The FBI's allegation that the tool replicated technology developed by Anthropic, a San Francisco AI safety company, raises questions that extend beyond any single procurement failure: how do nations protect the fruits of their ingenuity when the tools of replication travel faster than the rules meant to govern