Anthropic Accuses Alibaba of Largest AI Distillation Attack: 28.8M Fraudulent (yipzap.com)

Anthropic has lodged a serious accusation against Alibaba, claiming the Chinese tech giant orchestrated the largest AI model extraction, or distillation attack, in history, involving 28.8 million fraudulent exchanges with its AI model, Claude. This operation reportedly spanned 44 days, utilizing around 25,000 fake accounts to systematically target Anthropic's key capabilities, particularly in software engineering and agentic reasoning, which are central to its cutting-edge technologies. This level of sophistication not only highlights potential threats to proprietary technology but raises alarms about the safety and security of AI systems lacking proper safeguards. The implications of this attack extend beyond mere intellectual property theft, as Anthropic warns that undisciplined AI models could pose significant national security risks. This allegation comes amid heightened scrutiny of Chinese AI companies, further complicated by recent U.S. regulatory actions against Alibaba, including its designation as a Chinese military company. Following the announcement, Alibaba's stock plummeted, reflecting investor fears regarding compliance in the Chinese AI sector. As pressure mounts on policymakers, Anthropic calls for coordinated action between the government and industry to combat the growing threat of illicit AI distillation.