Global Power Plays

Anthropic hits back after China warns of Claude Code ‘backdoor’ risks

China advised users to uninstall Anthropic’s Claude Code over alleged 'backdoor' risks; Anthropic responded that the product was not intended for users in China and pointed to account and geofencing rules. The exchange highlights how states and tech firms exert control over cross-border AI access and may accelerate geofencing, removals, or regulatory clashes.

What happened

China warned people to uninstall Claude Code, a product from US AI company Anthropic. Beijing said the app could act like a "backdoor." Anthropic replied that users in China were not meant to be on the product.

Both sides framed the move as a safety step. Officials in Beijing pointed to national-security risks. Anthropic pointed to account rules and rights over its code.

Who wins here

Governments and big tech hold the most leverage. China gains control over what tools are allowed inside its borders. Anthropic and other US AI firms get to shape how rules are set by stressing control over their products.

Civil users lose ground. People in China may lose access to new tools. Outside users face louder politics around which AI firms can serve which countries.

How the play works

This is a mix of tech control and political pressure. China uses public warnings to force removals or block access. Firms push back by saying users broke terms or by pointing to tech protections.

The practical effect is fast. Apps get pulled or throttled. Firms change distribution, add geofencing, or sue for policy clarity.

Why it matters

Access to AI shapes job tools, news checks, and research. When a state blocks a tool, everyday people lose a way to work or learn. When firms limit service by country, it raises the cost of cross-border work.

The fight also sets rules for future AI moves. If nations can force cuts, companies must plan for political splits, not just bugs or abuse.

What to watch next

Watch for three things: tech changes from Anthropic, formal Chinese rules or enforcement steps, and similar warnings from other countries. Also watch whether firms add clearer geofencing or legal fights over user bans.

LensGlobal Power Plays
TypeReporting
PublishedJuly 9, 2026
Read time3 min read
SourceScmp
Where the facts come from

The facts in this story were first reported by Scmp. What you're reading here is our take on what it means for power and for you.

Read the original at Scmp
Related topics

More stories on these topics

AnthropicClaude CodeChinaAIapp blockinggeofencingSouth China Morning Postglobal-power-plays
Subscribe for moreExplore this lensBrowse all issues