Eight Ethical Stances
1. Attend to your own intentions, perceptions, and unease.
In working with AI, a subtle unease can surface: the tool is convenient, yet something feels unresolved; the output is useful, yet the authorship or judgment feels unsettled. That uncertainty does not always lead to an immediate answer, but it is a serious starting point for ethical reflection.
2. Begin with the person affected.
Begin with the person before you: their presence, feelings, and the possibility of thinking together. AI can imitate a particular person's voice or position, but it is not that person and should not replace a relationship with them.
3. Consider people beyond the immediate interaction.
AI use can affect people outside the visible situation, including those indirectly affected or affected later. A use that is convenient for one person may shift risk or burden elsewhere. A system that works well for many may still exclude a smaller group.
4. Read AI outputs alongside the systems behind them.
AI outputs should be read alongside the systems that produce them: data, institutions, goals, interfaces, and social conditions. To examine AI is also to examine the structures of society that make it possible.
5. Stay with the question of what is right.
Many questions do not have quick answers. Is it unfair to write text or make images with AI? What changes when people seek guidance from AI about personal decisions? The answer may change with time, context, and perspective. Treat uncertainty not as delay or weakness, but as evidence that ethical judgment is at work.
6. Verify independently.
AI outputs can sound natural and confident, but fluency is not evidence of accuracy. Treat AI outputs as claims to be checked against reliable sources, context, and accountable human judgment. Verification is also a way to avoid harming people with incorrect information.
7. Use AI in ways you can explain yourself.
AI cannot take responsibility. That is why decisions and actions shaped by AI need to remain explainable by people. Ask who you could explain this judgment to, who could correct it if it is wrong, and who could respond if someone is harmed. Do not let responsibility disappear into AI.
8. Continue making.
Ethical reflection should not stop us from making things; it should shape how they are made, tested, shared, and revised. Ideas do not reach others unless they are given form. Nothing can be examined, used, or improved if nothing is made.