Implications of the future use of machine learning in complex government decision-making in Australia

Abstract

The rapid growth in Artificial Intelligence and machine learning algorithms is currently outpacing legal and regulatory efforts in Australia. This paper analyses the use of machine learning algorithms in complex government decision-making. It considers whether the use of machine learning can comply with merits review, judicial review and the requirements of responsible government. This paper concludes that given the current state of technology, changes are needed to ensure that machine learning does not undermine democracy in Australia. It recommends amendments are made to the Freedom of Information Act 1982 (Cth) to reduce the number of grounds on which departments can deny Freedom of Information requests concerning the use of algorithms in decision-making. It also recommends that decision-makers update policies and legislation to ensure that if decisions need to be remade – they are remade by human decision-makers. The paper further argues that greater research is needed to ensure that an adequate statement of reasons can be generated by automated decision-making processes.

Keywords: Law and Technology, Artificial Intelligence, Government Decision-Making, Judicial Review