76.3% of Chinese respondents believe AI is a threat to privacy in an 8,000 person survey carried out by state-run TV network CCTV and Tencent Research
Frank Hersey / TechNode :
Context & Ripple Effects
In 2018, an 8,000-person survey run by state broadcaster CCTV and Tencent Research found 76.3% of Chinese respondents see AI as a threat to privacy — a striking result given both sponsors sit at the center of China's AI deployment. The finding opened a thread of Chinese public-opinion research that later coverage kept pulling on.
A year later, a survey of 6,154 people in China showed the concern had a concrete shape: 74% wanted the option to choose traditional ID methods over facial recognition, and 84% wanted to review or delete their data. By 2022, though, 76% of Chinese citizens said they trusted AI companies — versus 35% in France, the UK, and the US — and by 2026 reporting framed Chinese citizens as more optimistic about AI than Americans. The 2018 number is the anchor of that arc: privacy anxiety coexisting with, not preventing, broad trust.
First-order effects
- Tencent Research's own sponsor relationship is implicated: one of China's largest consumer-data companies published evidence that its home market overwhelmingly views AI as a privacy risk.
- Facial-recognition and identity-verification deployers in China are put on notice that the public wants opt-outs and data deletion rights, not just accuracy improvements.
Second-order effects
- Companies rolling out face-based ID services face pressure to offer traditional alternatives as a product feature, since three-quarters of surveyed users want that choice.
- State-run measurement of AI sentiment gives regulators a recurring evidence base for consent and data-control rules, raising compliance costs for consumer AI products.
Third-order effects
- If the pattern holds — high privacy-threat perception alongside high trust in AI firms — China's AI governance develops as a managed trade-off: firms keep deploying, while the state channels privacy demands into controlled rights like review-and-delete rather than adoption slowdowns.
- The divergence from Western survey results, where concern skews toward civilizational or societal risk, points toward regionally distinct AI social contracts shaping which products each market accepts.
The trend: Chinese public opinion on AI splits between sharp privacy anxiety and pragmatic trust in AI companies, and successive surveys are turning that tension into the input for state-managed governance.