Are LLMs pushing political narratives?

Описание: I tried asking DeepSeek to help me adjust the presentation of an article I had written -- not the content. What I received back instead was something with completely different information, including several talking points taken straight out of the Kremlin's propaganda playbook.

Journalist Christo Grozev explains his experience using DeepSeek and asks the question -- why would this have been the response from an AI model? Is there something more concerning here than just a confused LLM?

#christogrozev #fail #politics