WASHINGTON (dpa-AFX) - The White House Office of Science and Technology Policy is sounding the alarm about foreign entities, mainly from China, carrying out large-scale operations aimed at stealing capabilities from advanced American AI systems.
This information came from a memo by Michael Kratsios, who described these efforts as sophisticated tactics designed to dodge security measures and capture proprietary AI knowledge.
Kratsios mentioned that these operations utilize tens of thousands of fake accounts and advanced tools to stay under the radar while they systematically access sensitive information about how models behave. He referred to this as a coordinated effort to 'exploit American expertise and innovation' by copying the capabilities of top AI systems.
A major point of concern is a method called model distillation, which enables developers to transfer knowledge from big, complex AI models to smaller, more efficient ones. While this technique is common in the industry, US companies believe it's increasingly being misused to replicate proprietary systems without permission.
Notably, significant AI players like OpenAI and Anthropic have previously raised concerns about these activities. They have pointed to a Chinese startup called DeepSeek, claiming it has tried to replicate the performance of leading US models through extensive distillation efforts. Anthropic has also reported noticing organized campaigns aimed at extracting capabilities from its Claude AI system.
On top of intellectual property worries, officials are cautioning that these distilled models might carry security risks. Kratsios emphasized that such models can strip away built-in safeguards, which could remove protections intended to ensure accuracy, neutrality, and safe use.
This situation reflects the increasing rivalry between the United States and China in the global race for AI supremacy. The Trump administration has already taken steps to limit China's access to advanced AI chips and is now looking into further actions to combat unauthorized AI extraction efforts.
Proposed responses include sharing intelligence with US AI companies, enhancing collaboration between public and private sectors, and finding ways to hold foreign entities accountable. Officials are also working on developing best practices to help detect and prevent large-scale distillation campaigns.
These allegations spotlight the rising tensions at the crossroads of technology, national security, and global competition, as both governments and companies navigate the challenge of safeguarding advanced AI innovations.
Copyright(c) 2026 RTTNews.com. All Rights Reserved
Copyright RTT News/dpa-AFX
© 2026 AFX News
