Language models are not capable of generating biased text.

Bias in language models only affects certain applications.

Language models may reflect the biases and stereotypes present in their training data.

Language models are designed to generate human-like language.