One useful reflection exercise they got us doing on the AI Ethics, Regulation and Compliance programme at Saïd Business School, University of Oxford is to looking at the philosophical foundations of ethics and apply them to modern AI usage. I had come across these theories in my undergrad degree and found them a bit dull and abstract to be honest, but it was interesting to reconsider these ethical frameworks:
📜Deontology (Kant): The most important thing is following duty and universal rules.
🪶Consequentialism or Utilitarianism (Bentham): The most important thing is whatever produces the best overall outcomes, i.e. maximising happiness, reducing suffering.
🏛️Virtue ethics (Aristotle): The most important thing is moral character, i.e. what would a person with good character do in this situation?
How do these apply to AI? Well it’s interesting to read through the AI policies and values of some of the big tech companies and see a mix of commitments to rules, outcomes and moral character. They may be committed to creating protocols that monitor AI usage to prevent misuse (Unilever); embedding stringent ethical standards within the whole AI lifecycle (Amazon); designing systems to ensure AI serves human decision making, not to replace it (IBM); and ensuring AI aligns with the company’s vision of social good (Microsoft).
It’s all good and interesting to see organisations base their approach on these ethical frameworks, with different flavours according to their key stakeholders and customers – but these ethics still seem a little abstract. Are there any contemporary ethical approaches that might be useful to provide a more critical lens?
💥 What might a decolonial ethical lens on AI give us? Would it make us think more clearly about whose data, stories and labour is being extracted or imposed, and who gets to define the problem in the first place?
💥 What about applying a more holistic (and non-western) framework like Ubuntu, which believes that “a person is a person through other persons” – would this push us to ask what systematic impact AI has on interconnectedness, reciprocity and community voice?
💥 What if we took a justice-based or feminist ethical approach? Who gains real freedoms through AI and who loses them? Who gets harmed at the margins? Who is doing the hidden work? And do people have a voice to contest and redress injustice?
💥 What if we took Capability Approach (Amartya Sen) as an ethical framework, to judge policies and technologies by whether they expand or restrict people’s freedoms and capabilities, e.g. by what they are genuinely able to do to improve their lives?
Maybe these approaches are unlikely to be explicitly adopted by big corporations, but as individual consumers and entrepreneurs (also thinking of my coaching friends here) we are free to bring these more critical lenses to our use of AI and make our practice more ethically grounded to our values.
What do you think?