# China’s AI chiefs stay quiet on risk warnings

> As U.S. AI leaders call for caution, China’s tech sector is focused on control, cheaper models and real-world use.
- Title: China’s AI chiefs stay quiet on risk warnings · CNBC
- Summary: As U.S. AI leaders call for caution, China’s tech sector is focused on control, cheaper models and real-world use. China’s AI push is less about public…
- Keywords: china, ai risks, openai, anthropic, sam altman, dario amodei, jensen huang, beijing, ai governance, deepseek, qwen, openpie
- Source: CNBC — https://cnbc.com/2026/09/16/chinas-ai-leaders-keep-quiet-despite-us-publicity-on-tech-risks.html
- Author: Evelyn Cheng
- Published: 2026-09-16T04:01:59+00:00
- Read time: 1 min
- Topics: china, ai risks, openai, anthropic, sam altman, dario amodei, jensen huang, beijing, ai governance, deepseek, qwen, openpie
## U.S. warnings land flat
Top U.S. AI leaders warned about runaway risks, but China’s major AI companies stayed quiet. Z.ai, Moonshot, MiniMax, Alibaba and Tencent offered no matching public alarm.
## China hears publicity
Ray Von of OpenPie said caution makes sense, but U.S. warnings feel overplayed in China. His view: firms should spend less time talking and more time building responsibly.

> They should just do it.
## Safety becomes rivalry
Dario Amodei urged U.S. companies to keep their AI lead over China. Beijing’s foreign ministry called the U.S. comments fear-mongering, while state media framed them as geopolitical messaging.

> I don't think this is a coincidence
## Beijing sets guardrails
China released a new AI Safety Governance Framework during cybersecurity week. It calls for labeling AI content and faster ways to detect AI risks.
## Control came early
China restricted public access to ChatGPT-like tools until approvals arrived in summer 2023. Its regulator now lists registered models and reviews services that could shape public opinion.
## Some answers stay off limits
Chinese chatbots can compete with U.S. models on performance and price, but political censorship remains built in. DeepSeek refuses to answer about the Tiananmen crackdown.

> I'm not able to help with that.
## Applications beat AGI
Chinese companies have pushed commercialization while U.S. labs race toward superhuman AI. U.S. chip controls have made revenue-generating uses more urgent for China’s firms.
## Cheaper models spread
By early 2026, Chinese chatbots were fighting for users with big holiday promotions. Low-cost open-source models from China gained traction abroad, including in the U.S.
## Enterprises want payoff
OpenPie mostly uses older or smaller Qwen and DeepSeek models. Von said many companies see productivity gains, but few have seen fast profit growth or cost cuts.

> The application side, the enterprise side, hasn't seen a lot of result yet.
## Key takeaway

China’s AI push is less about public warnings and more about state control, commercial rollout and lower-cost applications.