Even ethical challenges can arise in the processes of AI. Arguably, for instance, there are ethical obligations to:
Explain how data processes work to those affected by them, and it is probably easier to explain simpler processes.
Ensure a person has ultimate control over decisions which affect people, and that might be easier to ensure in the case of simpler processes.
More dramatically, at some point we might think intelligent machines have consciousness and emotions, and if so, we might think that that they deserve ethical consideration, rather than just the natural agents or persons they affect.
The course underlines that all or almost all of the ethical issues raised by sophisticated data processes apply to AI, but that there are some issues raised by or relevant to AI which are not raised by or relevant to some sophisticated data processes. Although they said this is a topic saved for later lectures, so we will move on from here.
Going through common terms in ethical reasoning, we have:
Claim: an assertion or statement that is either true or false.
Normative claim: expresses a value judgment or opinion about how things should be. Discusses what ought to be, rather than what is.
Non-moral: discusses a rule that should be followed that is not based on morality.
Moral: expresses a value judgment about what is right or wrong, or what is good or bad, in a more individual and personal context.
Descriptive claim: describes something how something is, was, or will be. Does not make value judgments.
Ethical claim: discusses appropriate behavior within a specific context. Often based on established rules or guidelines within a community or profession.
Two of many very broad ethical theories are:
Consequentialism: focuses on the overall goodness of the outcomes of an action. Based on the idea that the better consequences an action produces, the better or righter that action is. Can lead to the position of ethical altruism, the claim that people have a moral duty to help others.
Rights-based ethics: prioritizes respecting the fundamental rights of individuals, viewing human dignity as the core principle. Emphasizes the idea that everyone has certain inalienable rights that should be upheld regardless of the situation.
A third category, virtue ethics, emphasizes the importance of moral character understood in terms of the presence or absence of virtues (e.g., courage, kindness, and fairness). Rooted in ancient Greek philosophy tied to Aristotle, it focuses on what it means to be a 'good person'.
Virtue ethics shifts the focus from what AI does to how AI behaves and what values it embodies. Key applications include designing virtuous AI agents, design trustworthy systems, and promote long-term flourishing aka sustainability in various fields.