Yes, AI models are biased.

All AI models are biased in some way. DeepSeek won't discuss particular topics not because of its own desires but because of its maker's desires and environment.

Biases also exist in other ways; they're created from the data used to build them, local governments' laws, the founders' beliefs, religion and the accuracy of information.

xAI might have a slightly more biased opinion about Elon Musk; Llama might be more biased about the Metaverse. It's natural, sometimes also intentional.

I don't mind any of that. You cannot escape a bias.

The problem comes from the fact that you don't know the boundaries of a model lie. If you knew the things DeepSeek spoke about, you could make a decision, if you could judge ChatGPT on the same thing you could make a comparison. But you can't.

Models can be biased; they're not your creation, and you don't have to use them. But you need to closely assess where the models sit within your own political and moral compasses and if they conflict or support the desires of the business.