Language models are not capable of generating biased text.

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

Bias in language models only affects certain applications.

Language models are designed to generate human-like language.