Anthropic recently published a threat intelligence report, openly accusing seven Chinese AI companies for the third time of engaging in 'industrial-scale distillation' of its Claude model. However, the legality of distillation—a neutral technology in itself—remains ambiguous, as it is unclear whether its application constitutes infringement. Moreover, the evidence cited by Anthropic has not been independently verified by a third party, complicating efforts to determine the legal nature of these allegations. Alibaba, one of the companies named in the report, has firmly denied any wrongdoing.

Previously, Anthropic had been embroiled in copyright disputes for training its models with books sourced from pirated platforms, eventually settling the case for $1.5 billion. On one hand, the company underscores the potential risks associated with AI and advocates for a slowdown in the industry's rapid development. On the other hand, it is actively pursuing an initial public offering (IPO) and accelerating its commercialization efforts.

Furthermore, Anthropic has declined to sign an open-source letter and has lobbied for restrictions on open-source models, citing 'security' concerns. This move appears to be an attempt to preserve its technological advantages, safeguard its commercial interests, and solidify its position within the industry. Its strategy of maintaining closed-source models with high-profit margins has put pressure on industry chain profits, sparking resistance from Silicon Valley companies.

Meanwhile, the technological gap between Chinese open-source models and their international counterparts has narrowed to approximately 2.7%. Chinese models are now being commercially exported to overseas cloud platforms, where they participate in revenue-sharing arrangements, effectively reversing their previous roles in the global market.