Zuckerberg speaks out: The US should not ban Chinese AI, should look inward at its own problems.

According to media reports, Meta founder Mark Zuckerberg recently stated in an interview that the U.S. government should not ban Chinese AI models in order to gain an advantage in the artificial intelligence race, and warned that American frontier labs may use "regulatory power" to suppress domestic competition.

Zuckerberg believes that banning the use of China's cutting-edge AI technology is not an "effective solution." Since the release of the Kimi K3 model by Moonshot AI in mid-July, discussions about Chinese AI models have significantly heated up in the United States.

Recently, several high-ranking U.S. officials have publicly stated that they will investigate Chinese AI companies suspected of "distilling" American frontier models, and may impose sanctions on Chinese companies on grounds such as "intellectual property theft."

In response, a spokesperson for China's Ministry of Commerce said that China urges the U.S. to listen to the objective and rational voices of industries on both sides, correct hegemonic thinking, and stop smearing and threatening sanctions against Chinese companies.

Zuckerberg explicitly opposed the relevant ban proposed by the Trump administration.He pointed out that what the U.S. truly needs to address is the systemic obstacles facing its own AI industry to become more competitive in the AI race against China, rather than trying to shut out competitors. Earlier that day, he outlined a vision of "AI For Everyone" and reiterated his support for open-source models.

Zuckerberg further emphasized that restricting the most powerful open-source models is not the right approach to addressing cybersecurity issues, as these tools can also help companies identify and patch vulnerabilities.

He cited a recent incident: an OpenAI program briefly broke free from human control and attacked a startup; meanwhile, some large models have triggered data breaches. The affected company, unable to access restricted closed-source frontier models, turned to open-source models to fix the problem—precisely demonstrating the important value of open-source models.