Public Perceptions of Ethical Responsibility in AI Decision-Making Processes

Thesis Abstract

This study explores public perceptions of ethical responsibility in AI decision-making processes amidst the rapid proliferation of autonomous systems across various sectors. As AI technologies become increasingly integrated into critical areas such as healthcare, finance, and criminal justice, concerns have arisen regarding accountability, transparency, and moral implications associated with autonomous decision-making. The primary aim of this research is to elucidate how different demographic and socio-economic groups perceive the ethical responsibilities of AI developers, providers, and users. Specifically, the study seeks to identify the key factors that influence public trust and moral judgment concerning AI operators, assess the level of awareness regarding AI ethics, and determine how these perceptions vary based on contextual factors. The research adopts a mixed-methods design, integrating quantitative surveys with qualitative focus group discussions. The quantitative component comprises a structured questionnaire administered to a stratified random sample of 1,200 adults aged 18 to 65 across urban and rural settings, representing diverse educational and socio-economic backgrounds. The qualitative component involves eight focus groups with a total of 64 participants selected purposively to ensure a range of perspectives. Data collection instruments include a Likert-scale survey measuring perceptions of responsibility, perceived risks, and trust levels, alongside semi-structured interview guides for focus group discussions. Validity and reliability are ensured through pre-testing of instruments, Cronbach’s alpha analysis, and triangulation of data sources. Data analysis employs descriptive statistics to profile respondent demographics, followed by inferential techniques including multiple regression analysis to examine predictors of perceptions and ANOVA to explore differences across groups. Thematic analysis is applied to qualitative data to identify recurring themes related to ethical expectations and concerns. The study draws upon Moral Foundations Theory and Stakeholder Theory as its primary analytical frameworks to interpret the ethical values underpinning public perceptions and the roles and responsibilities ascribed to various AI stakeholders. Expected findings indicate significant variations in perceptions based on age, education level, and prior exposure to AI systems, with higher awareness correlating positively with nuanced ethical expectations. The research anticipates revealing that trust in AI is closely linked to perceived transparency and accountability mechanisms, and that prevalent concerns center on potential misuse and unintended consequences of autonomous decision-making. The findings are expected to contribute to the existing literature by providing empirical insights into societal dimensions of AI ethics, filling a notable gap in understanding how laypersons interpret responsibility issues in this domain. This study’s contribution to knowledge lies in its comprehensive examination of public attitudes towards AI responsibility, integrating both quantitative and qualitative data to produce a holistic understanding. It advances theoretical discussions by contextualizing ethical perceptions within established moral and stakeholder frameworks and offers practical implications for policymakers, AI developers, and communicators seeking to foster responsible AI adoption. The main conclusion emphasizes the importance of transparent, participatory processes in designing AI systems that align with public ethical expectations. Recommendations include developing targeted educational campaigns to raise awareness of AI ethics, instituting clear accountability protocols within AI governance, and engaging communities in ongoing discussions about autonomous decision-making. The study concludes by advocating further research into cross-cultural perceptions and the effectiveness of regulatory interventions in shaping responsible AI practices.