Quantifying Cognitive Bias Induction in LLM-Generated Content

TL;DR

This paper investigates how large language models (LLMs) can introduce cognitive biases into generated content, affecting human decisions, and evaluates mitigation strategies to reduce such biases.

Contribution

It provides a comprehensive analysis of bias induction in LLMs across multiple tasks and assesses the effectiveness of various mitigation methods.

Findings

LLMs exhibit framing bias in 26.42% of cases

Hallucinations occur in 60.33% of post-knowledge questions

Humans are 32% more likely to buy a product after reading LLM-generated summaries

Abstract

Large language models (LLMs) are integrated into applications like shopping reviews, summarization, or medical diagnosis support, where their use affects human decisions. We investigate the extent to which LLMs expose users to biased content and demonstrate its effect on human decision-making. We assess five LLM families in summarization and news fact-checking tasks, evaluating the consistency of LLMs with their context and their tendency to hallucinate on a new self-updating dataset. Our findings show that LLMs expose users to content that changes the context's sentiment in 26.42% of cases (framing bias), hallucinate on 60.33% of post-knowledge-cutoff questions, and highlight context from earlier parts of the prompt (primacy bias) in 10.12% of cases, averaged across all tested models. We further find that humans are 32% more likely to purchase the same product after reading a summary of the review generated by an LLM rather than the original review. To address these issues, we evaluate 18 mitigation methods across three LLM families and find the effectiveness of targeted interventions.