The Use and Misuse of Counterfactuals in Ethical Machine Learning

The Use and Misuse of Counterfactuals in Ethical Machine Learning

Explore the complexities of counterfactuals in ethical machine learning through this 21-minute conference talk presented at FAccT 2021. Delve into the research conducted by A. Kasirzadeh and A. Smart as they examine the use and potential misuse of counterfactuals in fair machine learning systems. Begin with an introduction to the topic, followed by essential background information and key definitions. Engage in a thought-provoking discussion that covers various ontological views and identifies potential problems associated with counterfactual reasoning in AI ethics. Gain valuable insights into the challenges and considerations surrounding the application of counterfactuals in developing fair and ethical machine learning algorithms.