The race for supremacy in Artificial Intelligence (AI) has escalated into a heated confrontation, marked by allegations of intellectual property theft between China and the United States.
Accusations of Misuse of Models
China accused American AI corporations of using Chinese models in their training processes. This incident involves a technique known as model distillation and has intensified geopolitical tension between the two nations in the battle for AI dominance. Meanwhile, Reuters sources indicated that US authorities are investigating Chinese companies suspected of improperly using American technologies, placing intellectual property at the heart of this debate.
Focus on Distillation Technique
Model distillation is a technical procedure used in AI development to transfer knowledge from a more sophisticated system to another that is lighter and more efficient. China's Ministry of Commerce stated that US companies had applied this method to extract knowledge from Chinese models during research and training phases. In turn, US authorities maintain that certain Chinese companies may have exceeded legitimate use of the technology, using the process to replicate functionalities of models created in the US.
Key Points of Conflict
The main elements in dispute include the demarcation between machine learning and technological copying, the safeguarding of intellectual property in AI systems, the use of open models in developing new tools, and potential trade barriers related to technology.
Moonshot AI at the Center of Controversy
The discussion gained prominence after the launch of the Kimi K3 model, developed by the Chinese startup Moonshot AI, which attracted attention due to its remarkable performance in programming tasks. Michael Kratsios, Director of the White House Office of Science and Technology Policy, as reported by Reuters, stated that the US possessed information suggesting that Moonshot used the distillation of Anthropic's Claude Fable 5 model to create Kimi K3. Kratsios added that the company allegedly established an internal platform to execute large-scale distillation processes, employing various access methods to conceal its identity.
In response to the accusations, Moonshot vehemently denied any wrongdoing. In communication with the Chinese newspaper National Business Daily, the company argued that the progress achieved by Kimi K3 resulted from internal modifications to the system's architecture, not from a mere reproduction of American technology.
Sanctions and Government Response
Scott Bessent, US Treasury Secretary, warned that Chinese companies could be subject to sanctions or inclusion in the Entity List, a mechanism that restricts access to American technologies. Bessent emphasized: 'Open source is not an open season for American intellectual property.' Conversely, the Chinese government stated that it would adopt 'all necessary measures' to defend its interests if national companies were harmed.
Expanded Scope of Competition
Reuters also reported that Anthropic detected over 3.4 million interactions with its models linked to Moonshot. The company mentioned that hundreds of fake accounts were used to access features such as data analysis, programming, and tool usage. This case demonstrates that the rivalry for artificial intelligence has transcended the simple creation of the most advanced models, now encompassing training, infrastructure, data, and technological protection in the competitive landscape between China and the United States.



