Washington is intensifying its strategy regarding artificial intelligence in relation to China, as lawmakers advance new export control legislation and the Trump administration enhances collaboration with major American AI firms. These actions stem from growing concerns that Chinese technology companies are rapidly narrowing the existing technological gap.
Anthropic Calls for Extended Export Controls
A senior representative from Anthropic warned on Wednesday that while the United States currently maintains a substantial lead in advanced “frontier” AI models over its Chinese counterparts, unauthorized replication efforts pose significant risks. The company urged tighter export regulations to prevent foreign entities from circumventing current restrictions.
Tarun Chhabra, who serves as Anthropic’s head of national security policy, stated during a discussion at the Aspen Security Forum that the US holds an approximate lead of six to nine months in frontier AI models. However, he clarified that this figure represents an average when analyzing various models.
The Threat of Model Distillation
Chhabra characterized the methods used by Chinese firms—which involve unauthorized “distillation”—as being highly adversarial. He explained that distillation allows rivals to extract valuable intellectual property from advanced AI models. Chhabra warned that without this practice, American technological progress could have put the United States “a year to 18 months ahead.”
According to Anthropic’s official statements, virtually all of China’s leading research laboratories are currently engaged in distilling not only US-developed models but also those developed by other US peers. In response to this activity, Chhabra noted that the company is actively shutting down accounts facilitating distillation on a massive scale, reportedly amounting to millions of accounts every week.
Government and Legislative Efforts
The escalating concerns are prompting coordinated action from various levels of government. The White House has initiated a new program designed to synchronize cybersecurity vulnerability assessments with leading AI companies. Meanwhile, Congress is preparing legislation intended to further restrict China’s ability to obtain advanced semiconductors and specialized manufacturing equipment.
