How Different Groups Prioritize Ethical Values for Responsible AI

AI practitioners and the general public have significantly different priorities for responsible AI: The study found that AI practitioners consider responsible AI values as less important overall and emphasize a different set of values compared to a representative sample of 743 US adults, highlighting a potential gap between those building AI systems and those affected by them.

Women and Black respondents prioritize responsible AI values more highly: Self-identified women and Black survey respondents found responsible AI values more important than other demographic groups, suggesting that those with experiences of systemic inequality may have heightened concerns about how AI systems could perpetuate or amplify existing biases.

Political ideology, not personal discrimination experience, predicts fairness concerns: Surprisingly, more liberal-leaning participants were more likely to prioritize fairness in AI than participants who reported experiencing discrimination, suggesting that abstract political values may shape AI ethics priorities differently than direct personal harm.

There is no universal consensus on which values matter most: While private companies, public sector organizations, and academic groups have outlined overlapping ethical values for responsible AI, the study reveals that different groups within the population prioritize these values differently, raising questions about whose values should guide AI development.

Who defines responsible AI matters: The research emphasizes that the process of determining what "responsible AI" means should include diverse public perspectives, not just those of AI practitioners and established institutions, to ensure AI systems reflect the values of the broader population they affect.