US AI firm Anthropic enables Chinese-language options for Claude chatbot

US artificial intelligence company Anthropic has introduced Chinese-language options to its Claude chatbot in supported markets, even as it continues to block users from mainland China.

Checks by the South China Morning Post on Thursday confirmed that Claude’s web interface now lets users select simplified and traditional Chinese as their preferred on-screen language. China also briefly appeared as a billing country option before disappearing from the settings.

Users on X have shared screenshots showing UnionPay, China’s dominant bank card network, as a payment option at checkout. The SCMP was unable to independently confirm whether UnionPay payments were active.

Anthropic did not immediately respond to a request for comment on Thursday.

While the updates sparked speculation among some Chinese users that Anthropic might be preparing to ease access restrictions, analysts said the features were likely to target overseas Chinese communities rather than signalling any plan to enter the mainland market.

Poe Zhao, a China technology analyst and the founder of Hello China Tech, said the changes aimed to serve Chinese-speaking users in supported markets such as Singapore, Malaysia, Taiwan and North America.

“Offering a product in Chinese and serving users in China are two different decisions,” Zhao said. “Anthropic can do the first without the second.

“The official supported-region list is the best indicator of its strategy, and it has not changed.”

Su Lian Jye, chief analyst at Omdia, agreed that Anthropic appeared to be targeting the global Chinese diaspora rather than attempting to “warm up the Chinese market”, describing the localisation as a commercially significant step to enhance user experience.

The distinction highlights the balancing act facing Anthropic as it seeks to expand Claude’s global footprint while enforcing some of the industry’s strictest controls against access from mainland China.

The updates follow what appeared to be a recent tightening of Anthropic’s geographical restrictions. Last week, several Hong Kong users reported abrupt account suspensions after accessing Claude via virtual private networks (VPNs).

Mainland China, Hong Kong and Macau remain excluded from the San Francisco-based company’s official list of supported regions.

Last month, Anthropic accused Chinese AI developers, including DeepSeek and Moonshot AI, of conducting “industrial-scale” distillation using fraudulent accounts and proxy services to harvest Claude’s outputs and reasoning traces.

Anthropic CEO Dario Amodei has also repeatedly warned of the risks to the United States’ national security posed by China’s AI advances, urging Washington to maintain strict export controls on advanced chips and equipment.

Analysts expect Anthropic to carefully screen new account sign-ups to maintain its enforcement policies without loosening standards.

“The line between Chinese-speaking users overseas and users based in China will always be difficult to draw precisely, given VPNs, foreign payment cards and resold accounts,” Zhao said. “Anthropic relies on payment checks, IP detection and account reviews to manage it. I don’t expect this change to make those rules looser.”

Su added that Anthropic might shift towards tighter identity verification to distinguish between the two user groups, which could provoke further friction with Beijing.

Industry experts said the move highlighted the growing global influence of Chinese-speaking developers.