In conversation with Artificial Intelligence: aligning language models with human values