Can artificial intelligence (AI) make ethical decisions? Engineering and social science perspectives.

The increasing reliance on artificial intelligence (AI) to make decisions that affect everyday human life raises important questions about whether people believe AI is capable of moral agency-a feature of intelligent beings-and functions as an autonomous ethical decision maker.We propose that differences in such beliefs are in significant part a function of educational and professional pathways.Specifically, we hypothesize that individuals exposed to engineering-oriented disciplinary pathways, as compared to social scienceoriented pathways, are more likely to believe that AI is capable of a conscious capacity for moral reasoning, awareness, and responsibility for its decisions.To test this hypothesis, we first developed and validated an initial scale to measure such differences in belief and then examined these beliefs across a series of studies.We find that students in engineering programs, employees in engineering-related work sectors, and participants exposed to a problem-solving orientation stereotypical of an engineer are more likely to believe that AI is capable of being an autonomous ethical decision maker compared to their social science-oriented counterparts.We discuss the theoretical and practical implications of our findings for AI development and deployment and the associated importance of having diverse disciplinary perspectives represented to manage potential risks and rewards.