How Inclusively do LMs Perceive Social and Moral Norms?
This addresses the problem of biased LM judgments in decision-making systems for diverse human values, though it is incremental as it quantifies existing disparities without proposing a new solution.
The paper investigated how inclusively language models (LMs) perceive social and moral norms across demographic groups, finding notable disparities with younger, higher-income groups showing closer alignment to human responses, raising concerns about marginalized perspectives.
This paper discusses and contains offensive content. Language models (LMs) are used in decision-making systems and as interactive assistants. However, how well do these models making judgements align with the diversity of human values, particularly regarding social and moral norms? In this work, we investigate how inclusively LMs perceive norms across demographic groups (e.g., gender, age, and income). We prompt 11 LMs on rules-of-thumb (RoTs) and compare their outputs with the existing responses of 100 human annotators. We introduce the Absolute Distance Alignment Metric (ADA-Met) to quantify alignment on ordinal questions. We find notable disparities in LM responses, with younger, higher-income groups showing closer alignment, raising concerns about the representation of marginalized perspectives. Our findings highlight the importance of further efforts to make LMs more inclusive of diverse human values. The code and prompts are available on GitHub under the CC BY-NC 4.0 license.