FT says Anthropic is closing loopholes that allowed Chinese access to Claude

Financial Times reports that Anthropic is moving to close loopholes that allowed engineers in China to keep using Claude despite access restrictions. The story says engineers were still finding ways to use AI models even under stringent rules. It adds a concrete enforcement and export-control angle to the Anthropic model-access story.

Key details: Anthropic is closing access loopholes for Claude, The FT says engineers in China were still finding ways to use restricted models, The issue connects model access, compliance, and export-control enforcement.

Why it matters: Model-access controls only matter if labs can enforce them against workarounds in strategically sensitive markets.