The hidden way dictatorships are shaping what AI tells you
Authoritarian regimes may be unintentionally influencing AI chatbots like ChatGPT to reflect pro-regime perspectives. A recent study indicates that state-aligned media is significantly represented in training datasets for large language models. This raises concerns about the potential for these AI systems to propagate state propaganda without direct intervention from authoritarian governments.
- ▪Over a billion people use chatbots weekly, with ChatGPT alone having 900 million users.
- ▪The study found that 1.64 percent of CulturaX's Chinese language documents echoed state-aligned media.
- ▪State propaganda was found to be 41 times more prominent in training data than Chinese-language Wikipedia articles.
Opening excerpt (first ~120 words) tap to expand
PoliticsThe hidden way dictatorships are shaping what AI tells youAuthoritarian states may have accidentally brainwashed ChatGPT.by Eric LevitzMay 21, 2026, 10:30 AM UTCShareGiftChinese President Xi Jinping and Russian President Vladimir Putin are seen during the official welcoming ceremony and introduction of delegations ahead of a meeting at the Great Hall of the People in Beijing, China on May 20, 2026. Kremlin Press Service / Handout/Anadolu via Getty ImagesEric Levitz is a senior correspondent at Vox. He covers a wide range of political and policy issues with a special focus on questions that internally divide the American left and right.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Vox.
Discussion
0 commentsMore from Vox
