Today I asked ChatGPT to compare and contrast itself to DeepSeek, on the most prevalent biases in their models, why it’s the most prevalent, and how to mitigate those biases. ChatGPT's first comparison was political. I had to refine the question to tell ChatGPT to base its assessment solely on technology.