Humane Conversations: understanding and mitigating bias in automated AI systems