Anthropic has alerted US lawmakers to an alleged large-scale effort by Alibaba to illicitly access its Claude AI technology. In a letter dated June 10 sent to Senators Tim Scott and Elizabeth Warren of the US Senate Banking Committee, Anthropic claimed that operators linked to Alibaba utilized thousands of fraudulent accounts to gain access to its model. This disclosure follows a June 12 government mandate that forced Anthropic to restrict access to its advanced AI models, Mythos and Fable 5.

The company's correspondence highlights growing concerns regarding AI distillation, a process where a smaller model is trained to replicate the output and reasoning capabilities of a larger, more sophisticated AI. Security experts and government officials are increasingly focused on how such techniques might allow entities to bypass export controls or security guardrails by extracting knowledge from proprietary models. This incident has sharpened existing national security anxieties surrounding the protection of US-developed AI technologies from unauthorized foreign acquisition or manipulation. The Senate Banking Committee is currently investigating these developments as part of a broader review of AI security and access control policies.

Anthropic has alerted US lawmakers to an alleged large-scale effort by Alibaba to illicitly access its Claude AI technology. In a letter dated June 10 sent to Senators Tim Scott and Elizabeth Warren of the US Senate Banking Committee, Anthropic claimed that operators linked to Alibaba utilized thousands of fraudulent accounts to gain access to its model. This disclosure follows a June 12 government mandate that forced Anthropic to restrict access to its advanced AI models, Mythos and Fable 5.

The company's correspondence highlights growing concerns regarding AI distillation, a process where a smaller model is trained to replicate the output and reasoning capabilities of a larger, more sophisticated AI. Security experts and government officials are increasingly focused on how such techniques might allow entities to bypass export controls or security guardrails by extracting knowledge from proprietary models. This incident has sharpened existing national security anxieties surrounding the protection of US-developed AI technologies from unauthorized foreign acquisition or manipulation. The Senate Banking Committee is currently investigating these developments as part of a broader review of AI security and access control policies.