These are open questions! As you ask them, keep an open mind and be ready to go “off script.” Your students’ answers may surprise you.
So you want to have a conversation about GenAI ethics? If we do not all start from the same ethical premises, the risks of an over-committed shouting match—or an under-committed “well, it’s just up to everyone to decide for themselves”—are high.
Let’s start with the big question: What on Earth do we mean by “GenAI ethics”?
Meta-Ethics
Most conversations start with an assumption that there is a clear right/wrong answer to questions about GenAI impact. But it’s hard to assert absolute rules:
- Where do we see right and wrong in nature?
- If ethics are essentially theological, based in religion, who would purport to speak on behalf of the divine, and how should we trust them? Can we accept church pronouncements about artificial intelligence on faith?
- Sample Resource: Pope Leo has expressed skepticism about GenAI from a religious perspective.
- By what “gold standard” might we judge absolute rules about right and wrong? Who gets to adjudicate when people disagree?
In hedging responses that start out, “well, it depends…” we often find that the default meta-ethical stance is a kind of relativism: GenAI is good or bad insofar as it has beneficial or harmful effects on people, living things, or the environment as a whole.
Ethical and Moral Systems
If we accept that there are no clear, bright lines to give us immediate answers to these questions, we’ll have to decide which ethical system we want to use. In starting this kind of conversation, it may be useful to ask:
- How would you judge the impact of a technology on humans?
- During the Cold War, the USA and USSR imagined the outcomes of nuclear war, and invented terms like “megadeaths” to describe acceptable levels of destruction and death. Does this make you uncomfortable? How should we measure impact on human life?
- Are humans the only beneficiaries or victims of new technologies? How can we weigh human lives with or against other lives?
Most conversations about GenAI impact start from a utilitarian stance, gauging benefits and harms by their material results. But utilitarianism can invite us to accept terrible consequences in the name of great potential gains.
- What other systems might we use?
- How might we gauge GenAI effects on quality of life, rather than on material results?
- Many of those who design and market GenAI technology balance out any ill effects by saying that they are looking after the long term well-being of the human race—the untold billions of people yet to be born. Where does this reasoning invite fallacies or ethical problems?
- Sample Resource: Gebru and Torres (2025)