How People in China Keep Outsmarting Anthropic's Geolocation Restrictions (www.wired.com)

Anthropic's attempts to restrict access to its AI model, Claude, in China have been met with ingenuity from local users who have developed numerous workarounds. Despite strict measures, including account bans and identity verification processes aimed at preventing unauthorized access, Chinese startups and tech enthusiasts have leveraged tools like virtual private networks (VPNs) and underground markets to circumvent these barriers. This has led to the emergence of a black market for Claude accounts and a network of "transfer stations" that function as intermediaries, allowing users to access Claude's capabilities more easily and affordably. The significance of this situation is manifold for the AI/ML community, highlighting ongoing global disparities in AI access and the complexities of enforcing geographical restrictions in a hyper-connected digital landscape. Users in China favor Claude over local alternatives, such as DeepSeek and Z.ai, due to its superior performance, particularly in programming tasks. The situation raises critical questions regarding AI safety and ethics, as the reliance on unauthorized access routes can expose users to security threats and misuse of sensitive data. As demand for Claude persists despite restrictions, this cat-and-mouse dynamic poses ongoing challenges for both AI developers and policymakers aiming to maintain control over their technologies on a global scale.