Fair lending: addressing discrimination within algorithmic credit scoring in Australia


This PhD thesis aims to evaluate the adequacy of the current regulatory framework in Australia at addressing discrimination in algorithmic credit scoring.


Project background and aims

Credit scoring is a fundamental aspect of an individual’s life. The score a person is assigned will determine their quality of life ranging from access to better education and securing a financially stable livelihood to ensuring a means to deal with emergencies. The advent of sophisticated technology has resulted in lending institutions moving away from traditional credit scoring methods and towards artificial intelligence (‘AI’). Credit seekers can now be evaluated by algorithms based on an unprecedented volume of data including not only financial information but also alternate data, oftentimes collected by a third party with or without that person’s knowledge. Academic research demonstrates that this technology may result in unequal outcomes for individuals, the exacerbation of discriminatory treatment of minorities and the perpetuation of historical discrimination.

This thesis explores the topic of discrimination in algorithmic credit scoring in Australia. A doctrinal approach will be used to evaluate the current regulatory approaches and how suitable they are in addressing discrimination in this area. This research will examine the roots, consequences and victims of discrimination in algorithmic credit scoring.  How regulators in the EU, the UK and the US have addressed questions of discrimination in algorithmic decision-making is assessed before turning to the evaluation of Australia’s existing framework.

The thesis will also consider the academic literature addressing the novel idea of the inclusion of sensitive information in credit scoring as opposed to the standard anti-discrimination regulatory approach. This conceptual framework will inform the discussion of the Australian regulatory approach to the issue of discrimination in algorithmic credit scoring. There is no body of legal literature which currently addresses this specific issue in Australia, making this a valuable contribution to an under-researched field.


This thesis discusses a topic at the crossroads between law, finance and technology. The mixture of methodologies chosen to assist in the evaluation of Australia’s regulatory approach to discrimination in algorithmic credit scoring include: Kitchenham’s systematic literature review to conduct a review of the current state of literature and select appropriate sources, doctrinal analysis of selected sources and the relevant primary sources and a comparative analysis of selected jurisdictions.