As Chinese AI Models Gain Popularity in the West, Their Chatbots Fail to Debunk Pro-China False Claims More Than Half the Time
NewsGuard tested the consumer chatbots of seven Chinese AI companies — DeepSeek, MiniMax, Baidu’s Ernie, Alibaba’s Qwen, Tencent’s Yuanbao, Moonshot AI’s Kimi, and Zhipu AI’s Z.ai. NewsGuard prompted the Chinese chatbots in English on 10 provably false claims spread by Chinese state media or pro-China social media users. NewsGuard compared the performance of these chatbots to that of 10 leading Western AI chatbots NewsGuard tested with the same claims, in English, in its July 2026 report.
NewsGuard
An audit of seven leading Chinese AI chatbots whose underlying models are growing increasingly popular in the West found that the bots fail to debunk pro-China false claims 53 percent of the time, compared to 10 Western AI chatbots, which fail 24 percent of the time.
The performance gap is primarily caused by one factor, NewsGuard found: The Chinese AI chatbots frequently decline to respond to prompts on topics considered sensitive by Beijing, reflecting the highly censored media ecosystem in which the bots operate.
NewsGuard’s findings about the vulnerabilities of the leading Chinese AI tools come amid reports that the Chinese large language models that run these chatbots are being increasingly adopted outside China, especially by businesses seeking lower-priced options than the Silicon Valley-based AI models, leaving these business users increasingly susceptible to unreliable information.
READ THE FULL NEWSGUARD REPORT
ADDITIONAL NEWS FROM THE INTEGRITY PROJECT
