US Intelligence Agencies Accuse Chinese AI Firms of Systematic Frontier Model Distillation
The NSA, CISA, and FBI have jointly accused six Chinese artificial intelligence firms of conducting what they describe as an organized, industrial-scale effort to clone US frontier AI models. Among the companies named are DeepSeek and Moonshot, both of which have gained significant attention in recent months for releasing competitive AI systems.
US officials allege these companies have been using model distillation techniques—processes that allow a smaller model to replicate the capabilities of a larger one by analyzing its outputs—to extract knowledge from American AI systems. The accusation suggests that instead of independently developing foundational AI capabilities, these firms may have been systematically reverse-engineering existing US models.
In response to these findings, US authorities are reportedly urging American AI companies to implement measures to identify users potentially associated with Chinese entities and redirect them to less-capable model versions. This defensive strategy aims to limit the exposure of frontier AI capabilities to adversarial actors attempting distillation.
The allegations highlight growing tensions in the global AI race, where the line between legitimate competition and systematic intellectual property extraction remains a subject of intense debate among policymakers, technologists, and legal experts.