Chinese tech giant Alibaba has issued strict instructions to all its employees to immediately stop using artificial intelligence (AI) created by Anthropic. The policy, which takes effect on July 10, requires company staff to delete all related models and agent products, following the classification of the software as a high-risk category.
This decision was made following an escalation of tensions between the two companies. Previously, Anthropic filed an official complaint with the United States Senate accusing Alibaba of illegal attempts to extract their AI capabilities. Anthropic claims that the e-commerce giant engaged in large-scale distillation, a method of training AI models that uses the outputs of a more advanced model to enhance their own internal models.
In response to the new policy, Alibaba is now directing all its employees to switch to using the company's internal AI assistant, known as Qoder. This move is seen as Alibaba's effort to strengthen its technological independence amid tight restrictions on access to US-origin AI models for business entities in China.
This situation reflects the increasingly complex dynamics of global technology competition. Based on Anthropic's terms of service, companies from China and other countries on the US restriction list are indeed prohibited from using their AI models. However, reports reveal loopholes in access through third-party entities in Singapore or the use of virtual private networks (VPNs) by engineers at various major Chinese tech companies to access Claude services.
