Process- and Tool-centred Solutions for Ensuring Respect for Individuals’ Fundamental Rights in Algorithmic Credit Scoring

Abstract

Respect for fundamental rights is one of the values of the Rule of Law aimed at ensuring a society in which individuals can freely develop their identity, personality, and social relations and thus reach their potential to achieve set goals. With credit scoring using a subset of AI algorithms known as machine learning (ML) algorithms, which affects one’s access to credit, private life, and personal data, come threats to this value, particularly regarding the rights to non-discrimination, privacy, and data protection. Since access to credit determines individuals’ ability to fully participate in society or improve their standard of living, inequality in access due to errors or algorithmic discrimination cannot be treated solely as a matter of private law in any democracy upholding the Rule of Law and other EU values.

This thesis identifies gaps in EU legislation in regard to ensuring respect for said rights in algorithmic credit scoring and presents EU-level solutions, targeting both the process of and tool for this type of credit scoring. First analysed is the impact of the use of ML algorithms on the interpretation of the rights as protected in the EU. This is followed by an analysis of the risks algorithmic credit scoring poses to the respect for these rights and the extent to which relevant EU legislation can mitigate them, whereby the regulation of algorithmic credit scoring is considered in the contexts of consumer protection, data protection, and AI safety. Finally, solutions to fill the gaps in the legislation are proposed, with the conclusion acknowledging that considering the social and democratic importance of access to credit, the use of AI to assess how financially sound individuals are to justify the extension of credit may nevertheless not be a reality the EU should embrace.