# Three Kinds of AI Ethics

arXiv: https://arxiv.org/abs/2503.18842 · HTML: https://mlsanity.com/paper/2503.18842/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 38 references that are in the map.
- First posted on arXiv 24 Mar 2025.

## Builds on (most cited first, 8 of 8)

- [Artificial Intelligence, Values and Alignment](https://mlsanity.com/paper/2001.09768/) (2020)
- [Principles alone cannot guarantee ethical AI](https://mlsanity.com/paper/1906.06668/) (2019)
- [Beneficent Intelligence: A Capability Approach to Modeling Benefit, Assistance, and Associated Moral Failures through AI Systems](https://mlsanity.com/paper/2308.00868/) (2023)
- [From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices](https://mlsanity.com/paper/1905.06876/) (2019)
- [Machine Learning and Theory Ladenness -- A Phenomenological Account](https://mlsanity.com/paper/2409.11277/) (2024)
- [The Effectiveness of Embedded Values Analysis Modules in Computer Science Education: An Empirical Study](https://mlsanity.com/paper/2208.05453/) (2022)
- [What do we teach to engineering students: embedded ethics, morality, and politics](https://mlsanity.com/paper/2402.02831/) (2024)
- [A Capability Approach to AI Ethics](https://mlsanity.com/paper/2502.03469/) (2025)

## Abstract (arXiv)

There is an overwhelming abundance of works in AI Ethics. This growth is chaotic because of how sudden it is, its volume, and its multidisciplinary nature. This makes difficult to keep track of debates, and to systematically characterize goals, research questions, methods, and expertise required by AI ethicists. In this article, I show that the relation between AI and ethics can be characterized in at least three ways, which correspond to three well-represented kinds of AI ethics: ethics and AI; ethics in AI; ethics of AI. I elucidate the features of these three kinds of AI Ethics, characterize their research questions, and identify the kind of expertise that each kind needs. I also show how certain criticisms to AI ethics are misplaced, as being done from the point of view of one kind of AI ethics, to another kind with different goals. All in all, this work sheds light on the nature of AI ethics, and sets the groundwork for more informed discussions about the scope, methods, and training of AI ethicists.