Five minutes. What happened, and what to do about it. All stories from 4 October 2026
Free Gemini users lose the Flash and Pro models from 9 October
Free accounts drop to Flash-Lite. The $19.99 AI Pro plan gets Deep Think.
Chinese AI models give balanced answers on sensitive topics only 17 to 41 percent of the time, study finds
German AI firm Aleph Alpha tested models from Alibaba (Qwen), DeepSeek and Moonshot AI (Kimi) on 967 sensitive topics such as Tiananmen, Taiwan and Xinjiang. Its own AI scoring rated only 17 to 41 percent of answers as balanced, and the rest repeated state doctrine, deflected or refused. The slant also appeared in answers that did not mention China. Aleph Alpha sells sovereign AI to governments, so it has a commercial interest in the result.
- Find out whether any chatbot or writing tool you use is built on Qwen, DeepSeek or Kimi, and test it on a few questions close to your own market before it answers customers.
- For content that touches politics, history or human rights, have a person read the draft before it goes out, whichever model wrote it.
Source: The Decoder. This is our short summary of their reporting. The paragraphs marked "Our read" and the things to do are SeedFoundry's opinion. The summary is drafted with AI tools. Image: Aleph Alpha, from its own announcement. How we source and credit
These are short summaries of other people's reporting, with our own commentary. Each story names its source and links to the original. How we source and credit