AI: The only way is ethics

Rapid advances in machine learning technology have broadened both the capability and the ubiquity of AI systems. But how can we be sure that the machines are reaching fair conclusions and what safeguards should be put in place?