Anthropic Accuses Alibaba of 'Illicitly' Accessing AI Models
Anthropic has accused Alibaba of illicitly accessing its AI models, escalating tensions between the US and Chinese tech firms over intellectual property in artificial intelligence.
Background
Anthropic is a US AI startup founded by ex-OpenAI staff, best known for its Claude family of large language models (LLMs). Alibaba is the Chinese e-commerce and cloud computing giant. The accusation — that Alibaba systematically scraped or reverse-engineered Anthropic's proprietary models without authorization — sits at the intersection of two hot-button issues: the intensifying US-China tech rivalry (with AI as the key battleground) and the murky legal landscape around "web scraping" and model theft. Precedents include lawsuits by OpenAI and The New York Times against data scrapers, and US export controls aimed at stopping Chinese firms from accessing advanced American AI. This is not the first time a US AI company has accused a Chinese competitor of intellectual property theft, but it is notable because Anthropic claims it detected the intrusion from inside its own systems, not just from public behavior of Alibaba's models.