People have similar biases, it's just a lot easier to test an LLM for biases than it is to test humans so we are more aware of LLM biases.