Date: September 8, 2026 Main source: National Security Agency Link: https://www.nsa.gov/Press-Room/Press-Releases-Statements/Press-Release-View/Article/4592113/nsa-and-others-warn-china-based-ai-companies-are-distilling-us-frontier-ai-mode/ Additional sources: https://www.reuters.com/technology/us-accuses-chinese-ai-firms-industrial-scale-theft-ai-technology-2026-09-08/
Summary
The NSA, the FBI and CISA published a joint warning accusing Chinese companies of running distillation campaigns against US AI models. The document names companies such as DeepSeek, Moonshot AI and Alibaba, and argues that these practices aim to extract capabilities from advanced models to cut research and training costs. The agencies consider the activity to have commercial, cybersecurity and national security implications. China rejected the accusations and said its advances come from its own innovation.
Why it matters
Distillation is moving from an optimization technique to a point of geopolitical and intellectual property friction. This could lead to new access controls, API limits, regional restrictions and identity requirements for using advanced models. Companies that depend on multiple providers could face sudden changes in availability or terms of service.
Practical implications
- Keep abstractions between applications and model providers so you can switch APIs quickly.
- Review provider usage policies before using one model's outputs to train or improve another system.