Chinese State Media-Affiliated Account Singles Out Anthropic Over Data Privacy Risks
nashnova research
A CCTV-affiliated social media account publicly accused Anthropic of allowing discretionary data sharing with U.S. intelligence agencies. Timed to a U.S.–China trade-talk window, the rare move to name a top American AI company carries signal value well beyond a technical compliance debate.
What exactly did Anthropic change in its privacy policy?
The account "Yuyuan Tantian" (玉渊谭天) disclosed that Anthropic has revised its privacy policy 13 times since 2023 — an unusually high frequency.
Key changes include provisions to transfer user data from Canada, Brazil, South Korea, and the EU to U.S.-based storage.
The scope of data collection also expanded — from three categories to six between June 2024 and September 2025.
This means → Anthropic has been steadily widening its data boundaries, each revision small, but the cumulative shift is significant.
Why does "discretionary sharing" with intelligence agencies stand out?
The account noted that Anthropic's revised terms allow the company to decide on its own — without a legal process — whether to share user data with U.S. intelligence agencies.
In plain terms = no court order, no government subpoena required — the company can hand over data at its own discretion.
The account also cited an Anthropic report stating that, after accusing certain Chinese firms of model distillation — compressing a large model's capabilities into a smaller one — the company already shared related technical indicators with intelligence agencies.
This reflects a shift from passive compliance to active intelligence output in Anthropic's relationship with the U.S. intelligence apparatus.
What do the job postings reveal?
"Yuyuan Tantian" flagged that of Anthropic's three threat-intelligence job listings, two explicitly prioritize candidates with Mandarin or Russian language skills.
Those roles also require government or military intelligence analysis backgrounds and a U.S. top-secret security clearance.
This means → Anthropic is building a dedicated intelligence-analysis team focused on the China and Russia vectors — these are not ordinary security hires.
Why this timing?
The post landed at a sensitive moment — U.S.–China trade talks are being prepared, and a meeting between Trump and Xi is expected.
A state-media affiliate publicly naming a leading U.S. AI company at this juncture is itself a signal.
This reflects data security being folded into the negotiating leverage of U.S.–China tech competition, not treated as a mere compliance matter.
市场有风险，内容仅供研究参考，不构成投资建议。