Hong Kong users caught offguard as Anthropic tightens VPN access to Claude
US artificial intelligence company Anthropic appears to have stepped up enforcement of geographical restrictions on its Claude models in Hong Kong, with local users reporting abrupt account suspensions after accessing the service through virtual private networks (VPNs). A wave of complaints surfaced on social media platforms X and Reddit on Thursday, as users who had operated accounts for months…
Hong Kong residents have been caught off guard as Anthropic, an artificial intelligence company, appears to have increased scrutiny of VPN access to its Claude models. Local users on social media platforms X and Reddit reported sudden account suspensions after attempting to use the service through virtual private networks (VPNs).
A Reddit user explained they had been using NordVPN from Hong Kong for six months before receiving an email from Anthropic's safeguards team, stating that their account would be revoked due to "suspicious signals" linked to usage policy violations. Despite Anthropic not publicly announcing a new policy regarding VPNs, the sudden wave of suspensions indicates a deliberate tightening of existing regional restrictions.
Hong Kong is not among Anthropic's supported locations, which also exclude mainland China and Macau. Tech-savvy Hong Kong users had previously bypassed blocks by routing connections through supported markets like Singapore using VPNs and sometimes providing foreign payment details. Anthropic has not yet responded to inquiries about the abrupt enforcement measures.
The company's actions follow recent disagreements with Chinese AI developers. In a report last month, Anthropic accused seven Chinese AI labs of "industrial-scale" model distillation, the practice of harvesting outputs from Claude to train competing models. These labs, including Beijing-based Moonshot AI and Hangzhou-based DeepSeek, allegedly circumvented geographic blocks using thousands of fraudulent accounts and third-party proxy networks.
Moonshot alone is said to have routed over 23 million exchanges to Claude between May and July, disguising their activities through accounts appearing to originate from Singapore and Japan. Beijing has dismissed US allegations of AI distillation as unfounded. The geopolitical tensions surrounding AI development have escalated recently.
In a recent essay, Anthropic CEO Dario Amodei called for frontier labs to moderate the release of advanced systems, warning that a Chinese lead in AI would pose a "grave danger" to the world. He urged the US to maintain export controls on advanced semiconductors to China.
Written by urgent.news from SCMP Tech's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
