Submission in response to Australian Human Rights Commission Discussion Paper, 'Human Rights and Technology', April 2020

Abstract

This submission addresses the relationship between human rights law, ethics, and artificial intelligence. It is a submission in response to the Discussion Paper on Human Rights and Technology published by the Australian Human Rights Commission in 2019. It includes an extended discussion of the AHRC's proposal for legislation regarding the explainability of AI-informed decision making, which would make clear that, if an individual would have been entitled to an explanation of the decision were it not made using AI, the individual should be able to demand both a non-technical and a technical explanation of the AI-informed decision. The submission engages with research in both law and machine learning, arguing that the proposal raises as many questions as it answers, and raises a large number of questions of detail that will need to be addressed in operationalising any such requirement. The submission also addresses the AHRC's proposals around (1) the relationship between human rights law and ethics in the regulation of artificial intelligence, (2) 'AI-driven decision-making' as an object of regulation; (3) accountability, (4) transparency and (5) government use of AI.

Keywords: artificial intelligence and law, explainability, transparency, automated decision-making, human rights

JEL Classification: K10, K30