Trust and Biases in AI Systems