Is AI capable of making ethical decisions? Can an algorithm be trained to recognize potential ethical issues in its own decision-making process, even adjust avoid harmful outcomes?
Or does it come down to humans carefully curating unbiased and reliable datasets in order for Artificial Intelligence to produce the right outcome, and who defines "right?" Certainly AI models can be transparent -- updating their own understanding of ethical standards based on feedback, articulating reasoning behind decisions, flagging biased or unreliable data points from its training sets.
Whether in healthcare or hiring decisions, the more DATA we have, the better DECISIONS we can make. It's not so much a question of whether AI is ready to make unsupervised decisions. A better question is how much better AI can be than humans.