I feel like AI can go in any direction we can and farther. The more data an algorithm has to analyze the more it can narrow down a conclusion. Isn't that the same thing we do? We take the data of our experiences and accessible knowledge and sift through the relevant parts to apply to an idea or situation and act based on the output of that data. As with us, if the parameters are set to process moral or ethical ideologies along the way they are considered or ignored based on priority set within the parameters. Or am I misunderstanding your question?