AI Ethics: Explanation and the autonomy of affected parties

Many worry that an increased reliance on machine learning for training decision algorithms will make such procedures opaque. Even the experts involved in developing or employing such algorithms would have no grasp on why it makes the decisions that it does.

Building on research on the value of causal explanation, I argue that there is a distinct risk of undermining the autonomy of the individuals affected by such opaque algorithms. Answers to why-questions can be crucial to maintaining our autonomy and using opaque algorithms to make decisions that are central to our life plans thus threatens that autonomy.

Bram Vaassen is a Swedish Research Council post-doctoral fellow at Umeå University and a visiting research fellow at Rutgers University. Most of his research focuses on causation and explanation, and its applications in metaphysics, philosophy of mind, and philosophy of AI.