Alignment, bias, privacy and governance.
In AI ethics, bias refers to systematic prejudices or errors in machine learning systems that produce unfair or discriminatory outcomes for particular groups of people.