LLM Recommenders Produce Demographic Inequality Through Content Preferences

arXiv:2610.41854v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly being integrated into systems that rank and recommend online content. We examine how this shift can produce demographic inequality in whose voices are surfaced. Using 10,000 Twitter/X posts written by 1,000 survey respondents and linked to verified self-reported characteristics across nine demographic attributes, we conduct a controlled recommendation experiment across three LLM providers, six recommendation objectives, and five levels of contextual information. By varying whether models receive post text alone, platform metadata, or explicit author demographics, we isolate three pathways to unequal recommendation: preferences for demographically patterned content, reliance on unequal engagement signals, and direct use of disclosed identity. We find that demographic inequality emerges from text alone. Models consistently favour longer, more negative, and more opinionated posts, particularly those concerning news and social issues, producing systematic differences in the demographic composition of recommended authors. Platform metadata generally amplifies these inequalities, while explicitly disclosing authors' demographic characteristics produces little additional change. Recommendation objectives mainly affect the magnitude of these patterns: controversial framing intensifies both content and demographic skew, whereas popularity-oriented framing dampens it, although for some attributes - most notably age - the objective can also reverse which group is favoured. LLM-based recommend

LLM recommenders produce demographic inequality through content preferences.

The study shows LLM recommenders amplify demographic inequalities by favoring certain content patterns linked to social inequities, potentially marginalizing groups indirectly. This exacerbates existing inequalities in online visibility and voice without explicit targeting. Evidence from a controlled experiment using real user posts supports these implications, warranting concern though not extreme dystopian outcomes.

LLM-based recommenders, even without explicit demographic targeting, systematically favor certain content types (longer, more negative, opinionated) that are demographically patterned. This leads to demographic inequality by amplifying existing social inequalities through content and engagement signals. The study identifies specific pathways and consistent outcomes across different LLM providers and objectives.

The study shows that LLM-based recommenders can reproduce demographic inequality through content preferences and engagement signals. This has a materially concerning implication for human flourishing.

SOURCE     arXiv cs.CY (Computers and Society) (tier 1)
   ↓
DOCUMENT   3b3cd96d-4086-4ca9-b64e-d7a7ddda5a07
           https://arxiv.org/abs/2610.41854
   ↓
EVIDENCE   2 extracted excerpts
   ↓
MODEL RUN  5 runs, methodology 1.0.1
   ↓
SCORE      -70  (Strongly adverse)
   ↓
CONFIDENCE 84%