Loading market data...

Retired General Warns Closed AI Systems Risk National Security as China's Open Model Gains Ground

Retired General Warns Closed AI Systems Risk National Security as China's Open Model Gains Ground

Retired General James 'Spider' Marks has warned that closed artificial intelligence systems create national security risks, as China's open-weight AI model, Kimi K3, narrows the performance gap with US systems. The warning highlights a growing tension between the push for proprietary AI and the advantages of open development.

The General's Warning

Marks, a retired US Army general, said that closed AI systems are a national security risk. He argued that such systems could be exploited by adversaries because their inner workings are hidden from public scrutiny. The lack of transparency makes it harder to identify vulnerabilities or ensure the systems behave as intended.

China's Open-Weight Model

China's Kimi K3 is an open-weight AI model, meaning its parameters are publicly available for researchers and developers to study, modify, and build upon. According to recent benchmarks, Kimi K3 is closing the gap with leading US systems, demonstrating that open models can compete with proprietary ones. This development challenges the notion that only closed, tightly controlled systems can achieve top performance.

The rise of Kimi K3 underscores the rapid progress of Chinese AI, even as US companies keep their most advanced models closed. The open-weight approach allows for faster iteration and broader collaboration, but it also raises concerns about potential misuse. Marks' warning suggests that the US may need to reconsider its strategy of relying on closed systems to maintain a technological edge.

The debate over open versus closed AI is not new, but the emergence of a competitive open-weight model from China adds urgency. Policymakers are now grappling with how to balance the benefits of openness—such as innovation and transparency—against the risks of exposing sensitive technology to adversaries.

Marks' comments come as the US government considers new regulations for AI development. The question remains whether closed systems can be made secure enough to justify their limitations, or whether the US should embrace open models to keep pace with China.