Overview and Commentary of the Draft Modifications to Employment Regulations Regarding Automated-Decision Systems

Abstract

Recent years have seen an increase in the number of tools being offered to automatically screen and make decisions about the suitability of job applicants, including algorithmically judged video interviews, game-based assessments, and CV screeners. Although such technologies have the potential to streamline the job application process for both applicants and employers, a number of concerns have been raised about the use of these tools in the high-stakes context of recruitment, particularly in relation to issues such as potential bias, lack of transparency, and the need for additional governance of these systems. In response, a number of jurisdictions have proposed legislation to address the use of automated employment decision systems, with California proposing modifications to its employment regulation to address the use of these tools. In this article, we provide an overview of these proposed modifications, contextualising them within the larger artificial intelligence ethics debates and other similar legislation, as well as industry activity. We then identify some key themes in the legislation, namely i) ambiguity in terms of providing and justifying discrimination based on protected characteristics; ii) the lack of penalties for non-compliance; and iii) the lack of active obligations required of employers by the proposed amendments.

Keywords: : equal opportunity, automated decision system, psychometrics, AI ethics, automated decision system