Sari la conținut

Română · beta. Traducerea a fost făcută automat și este încă în curs de revizuire. Citește pagina în engleză

perspectiva aiqbenchSiguranță AI

analiza aiqbench

This study investigates how Large Language Models systematically marginalize queer identities by generating narrow, stereotyped narratives rather than reflecting the full complexity of human experience. The research identifies specific failure modes—including harmful representations and discursive othering—that limit the utility of current AI systems for inclusive discourse. These findings are critical for developers aiming to align model outputs with diverse social realities and mitigate bias in generative applications.
Surse care acoperăarXiv, LessWrong
Ce este aceastaLectura unui model asupra articolului și a acoperirii conexe, scrisă pentru aiqbench. Nu sunt cuvintele publicației.

Unequal Voices: How LLMs Construct Constrained Queer Narratives

arXiv1 min de cititArticolul și perspectiva sunt în engleză.

arXiv:2507.15585v2 Announce Type: replace-cross Abstract: One way social groups are marginalized in discourse is that the narratives told about them often default to a narrow, stereotyped range of topics. In contrast, default groups are allowed the full complexity of human existence. We describe the constrained representations of queer people in LLM generations in terms of harmful representations, narrow representations, and discursive othering and formulate hypotheses to test for these phenomena. Our results show that LLMs are significantly limited in their portrayals of queer personas.

Abstract:One way social groups are marginalized in discourse is that the narratives told about them often default to a narrow, stereotyped range of topics. In contrast, default groups are allowed the full complexity of human existence. We describe the constrained representations of queer people in LLM generations in terms of harmful representations, narrow representations, and discursive othering and formulate hypotheses to test for these phenomena. Our results show that LLMs are significantly limited in their portrayals of queer personas.

Computers and Society (cs.CY); Artificial Intelligence (cs.AI)

arXiv:2507.15585 [cs.CY]

(or arXiv:2507.15585v2 [cs.CY] for this version)

https://doi.org/10.48550/arXiv.2507.15585 Focus to learn more arXiv-issued DOI via DataCite

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Text integral sub licența deschisă a publicației · 175 cuvinte

Subiecte conexe