Left, Right, or Center? Evaluating LLM Framing in News Classification and Generation

TL;DR

This paper investigates how large language models (LLMs) influence political framing in news summarization and generation, revealing a tendency toward centrist bias and comparing the ideological expressiveness of various models.

Contribution

It introduces a comprehensive evaluation of LLMs' framing biases in news tasks, highlighting systematic centrist tendencies and comparing model-specific ideological expressiveness.

Findings

LLMs tend to produce centrist summaries and classifications.

Grok 4 shows the highest ideological expressiveness among models.

Claude Sonnet 4.5 and Llama 3.1 demonstrate strong bias-rating performance.

Abstract

Large Language Model (LLM) based summarization and text generation are increasingly used for producing and rewriting text, raising concerns about political framing in journalism where subtle wording choices can shape interpretation. Across nine state-of-the-art LLMs, we study political framing by testing whether LLMs' classification-based bias signals align with framing behavior in their generated summaries. We first compare few-shot ideology predictions against LEFT/CENTER/RIGHT labels. We then generate "steered" summaries under FAITHFUL, CENTRIST, LEFT, and RIGHT prompts, and score all outputs using a single fixed ideology evaluator. We find pervasive ideological center-collapse in both article-level ratings and generated text, indicating a systematic tendency toward centrist framing. Among evaluated models, Grok 4 is by far the most ideologically expressive generator, while Claude Sonnet 4.5 and Llama 3.1 achieve the strongest bias-rating performance among commercial and open-weight models, respectively.