Anthropic Accuses Alibaba of Massive AI Model Extraction Campaign Amid Rising U.S.-China Tech Tensions

Anthropic has accused Chinese technology giant Alibaba of carrying out a large-scale effort to extract and replicate its artificial intelligence capabilities, escalating tensions over AI competition between the United States and China.

In a letter sent to the U.S. Senate Committee on Banking, Housing, and Urban Affairs, Anthropic claimed that Alibaba orchestrated what it described as “the largest known distillation attack” against the company to date.

The letter, addressed to Senator Tim Scott and Senator Elizabeth Warren on June 10, alleged that operators affiliated with Alibaba and its AI research division conducted approximately 28.8 million interactions with Anthropic’s models between April 22 and June 5.

According to Anthropic, the activity was carried out using roughly 25,000 fraudulent accounts and was designed to facilitate model distillation, a process in which outputs from a more advanced AI system are used to train a smaller and less capable model.

Distillation has become a controversial topic within the AI industry as developers seek ways to protect proprietary models from being replicated by competitors.

Anthropic argued that the scale and sophistication of the activity represented a significant threat to U.S. AI leadership and called for stronger coordination between industry and government to combat unauthorized model extraction.

“We believe combating the threat of illicit distillation requires coordinated action between government and industry, and we will continue working with Congress and the Administration to maintain American AI leadership,” an Anthropic spokesperson said.

Alibaba has not publicly responded to the allegations.

The accusations come amid increasing scrutiny of AI model security and intellectual property protection. Earlier this year, the White House Office of Science and Technology Policy issued guidance aimed at helping AI companies detect and respond to industrial-scale distillation efforts.

Anthropic said Alibaba’s alleged actions occurred despite those warnings from the U.S. government.

The company has previously raised concerns about similar activities. In February, Anthropic disclosed what it described as three separate industrial-scale distillation campaigns linked to Chinese AI firms DeepSeek, Moonshot, and MiniMax.

At the time, Anthropic warned that such operations were becoming increasingly sophisticated and urged closer collaboration among AI developers, cloud providers, and policymakers to protect advanced models.

The latest dispute unfolds against a backdrop of broader geopolitical competition over artificial intelligence, semiconductors, and advanced technologies. Washington has increasingly viewed leading AI models as strategically important assets, prompting tighter controls on technology exports and access.

Anthropic itself has recently become entangled in those policy debates. Earlier this month, the company revealed that it received an export control directive from the Trump administration requiring it to suspend access to its latest AI models, Claude Fable 5 and Claude Mythos 5, for foreign nationals worldwide, including some foreign employees working within the company.

According to Anthropic, the directive cited national security concerns but did not specify the exact rationale behind the restrictions.

The company said senior executives subsequently traveled to Washington to meet with administration officials in an effort to resolve the matter. Discussions remain ongoing, and Anthropic has not yet provided a timeline for when access to the affected models may be restored.

The allegations against Alibaba underscore the increasingly strategic nature of artificial intelligence development, where access to cutting-edge models, training techniques, and intellectual property has become a central battleground in the global technology race. As governments and companies seek to secure competitive advantages in AI, disputes over model replication and technology transfer are likely to become more frequent and politically sensitive.