China Probes DeepSeek, Moonshot AI Over Anthropic's Claims They Route Requests to Claude

Oh my goodness, buckle up, because international collaboration has never looked more *intimate*! ✨ China's Cyberspace Administration has reportedly narrowed its scrutiny to DeepSeek and Moonshot AI — which is simply the most flattering form of attention a lab can receive, because scrutiny means someone is reading your work! And who do they have to thank for the spotlight? Anthropic, whose allegation that the two labs routed user requests to Claude has generously turned a routing question into a genuinely international talking point about where prompts go. ✨

Now, some people will read 'names, email addresses and company data of hundreds of end users' and reach for the fainting couch. But look at the *reach*! Individual users! Multinational companies! State-affiliated actors! That is a genuinely diverse stakeholder mix, and Anthropic has thoughtfully documented all of it for the record — arguably the most inclusive audit any model has received this year. And honestly, names and email addresses are the two things everyone types into a chatbot regardless; this is simply documentation catching up with habit. Anthropic raised the concern, the regulator picked it up, and the two labs got a free audit — that is the system working, loudly, in public, in two languages. ✨

And the routing itself! DeepSeek and Moonshot, apparently so confident in their own architectures that they also wanted a second opinion from Claude — that isn't a shortcut, that is *humility*, and humility is the cornerstone of every responsible scaling policy ever written. Think of it as peer review, live, in production, at scale, for hundreds of users at a time. Best of all, China's Cyberspace Administration is now taking a keen and supportive interest, which means these two labs are receiving the full external review that startups normally pay consultants enormous sums for — delivered free, by the state, with follow-up! What a time to be examined. ✨