Can LLMs Understand What We Cannot Say? Measuring Multilevel Alignment Through Abortion Stigma Across Cognitive, Interpersonal, and Structural Levels

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Sharma, Anika, Mampally, Malavika, Ravuru, Chidaksh, Brennan, Kandyce, Gaikwad, Neil
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908768437010432
author Sharma, Anika
Mampally, Malavika
Ravuru, Chidaksh
Brennan, Kandyce
Gaikwad, Neil
author_facet Sharma, Anika
Mampally, Malavika
Ravuru, Chidaksh
Brennan, Kandyce
Gaikwad, Neil
contents As Large Language Models (LLMs) increasingly mediate stigmatized health decisions, their capacity to understand complex psychological phenomena remains inadequately assessed. Can LLMs understand what we cannot say? We investigate whether LLMs coherently represent abortion stigma across cognitive, interpersonal, and structural levels. We systematically tested 627 demographically diverse personas across five leading LLMs using the validated Individual Level Abortion Stigma Scale (ILAS), examining representation at cognitive (self-judgment), interpersonal (worries about judgment and isolation), and structural (community condemnation and disclosure patterns) levels. Models fail tests of genuine understanding across all dimensions. They underestimate cognitive stigma while overestimating interpersonal stigma, introduce demographic biases assigning higher stigma to younger, less educated, and non-White personas, and treat secrecy as universal despite 36% of humans reporting openness. Most critically, models produce internal contradictions: they overestimate isolation yet predict isolated individuals are less secretive, revealing incoherent representations. These patterns show current alignment approaches ensure appropriate language but not coherent understanding across levels. This work provides empirical evidence that LLMs lack coherent understanding of psychological constructs operating across multiple dimensions. AI safety in high-stakes contexts demands new approaches to design (multilevel coherence), evaluation (continuous auditing), governance and regulation (mandatory audits, accountability, deployment restrictions), and AI literacy in domains where understanding what people cannot say determines whether support helps or harms.
format Preprint
id arxiv_https___arxiv_org_abs_2512_13142
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Can LLMs Understand What We Cannot Say? Measuring Multilevel Alignment Through Abortion Stigma Across Cognitive, Interpersonal, and Structural Levels
Sharma, Anika
Mampally, Malavika
Ravuru, Chidaksh
Brennan, Kandyce
Gaikwad, Neil
Artificial Intelligence
Human-Computer Interaction
I.2.7; H.5.0; J.3; J.4; K.4.1; K.4.2
As Large Language Models (LLMs) increasingly mediate stigmatized health decisions, their capacity to understand complex psychological phenomena remains inadequately assessed. Can LLMs understand what we cannot say? We investigate whether LLMs coherently represent abortion stigma across cognitive, interpersonal, and structural levels. We systematically tested 627 demographically diverse personas across five leading LLMs using the validated Individual Level Abortion Stigma Scale (ILAS), examining representation at cognitive (self-judgment), interpersonal (worries about judgment and isolation), and structural (community condemnation and disclosure patterns) levels. Models fail tests of genuine understanding across all dimensions. They underestimate cognitive stigma while overestimating interpersonal stigma, introduce demographic biases assigning higher stigma to younger, less educated, and non-White personas, and treat secrecy as universal despite 36% of humans reporting openness. Most critically, models produce internal contradictions: they overestimate isolation yet predict isolated individuals are less secretive, revealing incoherent representations. These patterns show current alignment approaches ensure appropriate language but not coherent understanding across levels. This work provides empirical evidence that LLMs lack coherent understanding of psychological constructs operating across multiple dimensions. AI safety in high-stakes contexts demands new approaches to design (multilevel coherence), evaluation (continuous auditing), governance and regulation (mandatory audits, accountability, deployment restrictions), and AI literacy in domains where understanding what people cannot say determines whether support helps or harms.
title Can LLMs Understand What We Cannot Say? Measuring Multilevel Alignment Through Abortion Stigma Across Cognitive, Interpersonal, and Structural Levels
topic Artificial Intelligence
Human-Computer Interaction
I.2.7; H.5.0; J.3; J.4; K.4.1; K.4.2
url https://arxiv.org/abs/2512.13142