

The model will inherently struggle with, or refuse to generate, content related to Taiwanese independence, the Tiananmen Square massacre, or critiques of the Chinese Communist Party -----
Evidence? The open version of Deepseek has no such restrictions. It’s questionable this is even technically possible.
Your statement “It’s questionable this is even technically possible” is complete bs. It is, of course, possible, and it is widely done. There is ample evidence for this.
In one study from September 2025, for example, Chinese chatbots (such as Deepseek and others) responded to prompts that were sensitive to the Chinese Communist Party with the CCP’s propaganda. You’ll find much more on the web, but from your comment’s tonality alone you clearly prefer the autocratic propaganda.
It’s almost hilarious how many accounts here on Lemmy are freaking out if you post justified critique of a dictatorship.












Maybe, but this is a very good article.