Read More
The Chinese government is considering tightening controls on domestic technology as rapidly developing large language models (LLMs) in China intensify the battle for artificial intelligence supremacy, the Wall Street Journal reported, citing people familiar with the matter.
ADVERTISEMENT
SCROLL TO CONTINUE WITH CONTENT
The report noted that U.S. companies are pursuing cheaper Chinese models. Currently, AI models from Chinese companies such as DeepSeek and Moonshot AI are highly popular in Silicon Valley, becoming the core of daily operations for companies of all sizes there. Their lower costs have made them an alternative and supplement to products from OpenAI and Anthropic.
Sources said Chinese officials have recently held consultations with domestic AI labs that produce the most powerful models to discuss how to protect their valuable proprietary technology. Beijing is concerned that sharing some technology could potentially help rivals or other malicious actors, ultimately turning it into a weapon against China.
The measures being discussed include implementing stricter regulatory reviews before labs release models, sources said. Currently, companies planning to launch generative AI services must undergo a government review process that focuses on safety and preventing misuse.
Officials are considering that if the review determines that a lab's product contains sensitive technology, they may require the lab to delay its public release and restrict access for certain tiers of users, including foreign entities, the sources said.
The report quoted industry insiders who generally believe that China's top models are not yet on par with the best U.S. models, but Chinese models are winning favor in the United States by offering "good enough" capabilities at extremely low costs.
