Ethical Analysis in the Development of AI Systems that Automate Decision Making

Abstract

The development of artificial intelligence (AI) systems capable of making automatic decisions raises complex ethical challenges that require thorough examination. This study aims to explore how ethical principles such as fairness, transparency, responsibility, and accountability are operationalized in the AI development process. Using a qualitative approach with a multiple case study design, data were collected through literature review, semi-structured interviews with developers, ethicists, and policymakers, as well as analysis of internal policy documents and algorithm audits. The findings indicate that although there is awareness of the need to apply ethical principles, the integration of ethics into the technology development cycle remains incomplete. Key obstacles include the lack of harmonized audit standards, limited organizational capacity to foster an ethical culture, and insufficient technical tools to translate abstract principles into concrete practice. Technical approaches, such as the use of synthetic data and the “Ethics as a Service” governance model, have proven effective in enhancing fairness and accountability, yet their success depends heavily on institutional support and ongoing evaluation. This study recommends the development of a contextual and adaptive ethical framework for AI advancement in Indonesia.