Anthropic moves to block Chinese workarounds to access Claude

In short: Anthropic is trying to close access gaps after reports that some Chinese companies used workarounds to use its Claude AI tools despite a ban.

What happened

Anthropic, a US company behind the Claude AI assistant, is moving to shut down “loopholes” that let some Chinese companies keep using its products, according to people familiar with the matter.

The Financial Times reported that companies including Ant Financial accessed Anthropic tools such as Claude Code through workarounds. These included using cloud providers and overseas subsidiaries. In Ant’s case, employees reportedly used corporate Claude accounts through an intranet connected to a Singapore-based entity.

The report also said some ByteDance engineers have used personal Claude subscriptions and claimed reimbursements from the company. Engineers reportedly accessed those subscriptions using VPNs, which are tools that can make your internet connection look like it is coming from another country (like taking a detour to enter a building through a different door).

Anthropic has one of the strictest policies among US AI companies on use in China. It requires user verification and blocks payments from Chinese banks. The FT report said these workarounds do not appear to break US or Chinese law, but they do violate Anthropic’s terms of service, which ban Chinese companies and foreign entities owned by them from using its AI models.

Why it matters

Many everyday products now depend on the same AI tools that help write code, answer customer support questions, and analyze documents. When access rules tighten, companies may switch to other providers, and that can affect cost, quality, and which countries’ tech is used behind the scenes.

Source: Financial Times