Polarization by Default: Auditing Recommendation Bias in LLM-Based Content Curation

arXiv:2604.15937v1 Announce Type: cross
Abstract: Large Language Models (LLMs) are increasingly deployed to curate and rank human-created content, yet the nature and structure of their biases in these tasks remains poorly understood: which biases are robust across providers and platforms, and which can be mitigated through prompt design. We present a controlled simulation study mapping content selection biases across three major LLM providers (OpenAI, Anthropic, Google) on real social media datasets from Twitter/X, Bluesky, and Reddit, using six prompting strategies (textitgeneral, textitpopular, textitengaging, textitinformative, textitcontroversial, textitneutral). Through 540,000 simulated top-10 selections from pools of 100 posts across 54 experimental conditions, we find that biases differ substantially in how structural and how prompt-sensitive they are. Polarization is amplified across all configurations, toxicity handling shows a strong inversion between engagement- and information-focused prompts, and sentiment biases are predominantly negative. Provider comparisons reveal distinct trade-offs: GPT-4o Mini shows the most consistent behavior across prompts; Claude and Gemini exhibit high adaptivity in toxicity handling; Gemini shows the strongest negative sentiment preference. On Twitter/X, where author demographics can be inferred from profile bios, political leaning bias is the clearest demographic signal: left-leaning authors are systematically over-represented despite right-leaning authors forming the pool plurality in the dataset, and this pattern largely persists across prompts.