# Bias Amplification: Large Language Models as Increasingly Biased Media

arXiv: https://arxiv.org/abs/2410.15234 · HTML: https://mlsanity.com/paper/2410.15234/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- Placed in LLM synthetic data generation, one of 13 directions in Synthetic data and self-improvement for LLMs.
- Builds on 25 of its 49 references that are in the map.
- First posted on arXiv 19 Oct 2024.

## Placed in

- Synthetic data and self-improvement for LLMs / [LLM synthetic data generation](https://mlsanity.com/synth/llm-synthetic-data-generation/)

## Builds on (most cited first, 25 of 25)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [A Survey on Bias and Fairness in Machine Learning](https://mlsanity.com/paper/1908.09635/) (2019)
- [BBQ: A Hand-Built Bias Benchmark for Question Answering](https://mlsanity.com/paper/2110.08193/) (2021)
- [Men Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level Constraints](https://mlsanity.com/paper/1707.09457/) (2017)
- [Self-Consuming Generative Models Go MAD](https://mlsanity.com/paper/2307.01850/) (2023)
- [A Tale of Tails: Model Collapse as a Change of Scaling Laws](https://mlsanity.com/paper/2402.07043/) (2024)
- [A Note on Shumailov et al. (2024): `AI Models Collapse When Trained on Recursively Generated Data'](https://mlsanity.com/paper/2410.12954/) (2024)
- [The Curious Decline of Linguistic Diversity: Training Language Models on Synthetic Text](https://mlsanity.com/paper/2311.09807/) (2023)
- [Measuring Political Bias in Large Language Models: What Is Said and How It Is Said](https://mlsanity.com/paper/2403.18932/) (2024)
- [Model Collapse Demystified: The Case of Regression](https://mlsanity.com/paper/2402.07712/) (2024)
- [How Bad is Training on Synthetic Data? A Statistical Analysis of Language Model Collapse](https://mlsanity.com/paper/2404.05090/) (2024)
- [Data Feedback Loops: Model-driven Amplification of Dataset Biases](https://mlsanity.com/paper/2209.03942/) (2022)
- [Fairness Feedback Loops: Training on Synthetic Data Amplifies Bias](https://mlsanity.com/paper/2403.07857/) (2024)
- [Self-Consuming Generative Models with Curated Data Provably Optimize Human Preferences](https://mlsanity.com/paper/2407.09499/) (2024)
- [OpinionGPT: Modelling Explicit Biases in Instruction-Tuned LLMs](https://mlsanity.com/paper/2309.03876/) (2023)
- [Detecting Mode Collapse in Language Models via Narration](https://mlsanity.com/paper/2402.04477/) (2024)
- [Would Deep Generative Models Amplify Bias in Future Models?](https://mlsanity.com/paper/2404.03242/) (2024)
- [JobFair: A Framework for Benchmarking Gender Hiring Bias in Large Language Models](https://mlsanity.com/paper/2406.15484/) (2024)
- [Towards Auditing Large Language Models: Improving Text-based Stereotype Detection](https://mlsanity.com/paper/2311.14126/) (2023)
- [Envisioning the Applications and Implications of Generative AI for News Media](https://mlsanity.com/paper/2402.18835/) (2024)
- [SAGED: A Holistic Bias-Benchmarking Pipeline for Language Models with Customisable Fairness Calibration](https://mlsanity.com/paper/2409.11149/) (2024)
- [HEARTS: A Holistic Framework for Explainable, Sustainable and Robust Text Stereotype Detection](https://mlsanity.com/paper/2409.11579/) (2024)
- [Exploring the Escalation of Source Bias in User, Data, and Recommender System Feedback Loop](https://mlsanity.com/paper/2405.17998/) (2024)

## Built on by (most cited first, 12 of 12)

- [The Homogenizing Effect of Large Language Models on Human Expression and Thought](https://mlsanity.com/paper/2508.01491/) (2025)
- [From Five Dimensions to Many: Large Language Models as Precise and Interpretable Psychological Profilers](https://mlsanity.com/paper/2511.03235/) (2025)
- [Subliminal Transfer of Unsafe Behaviors in AI Agent Distillation](https://mlsanity.com/paper/2604.15559/) (2026)
- [Synthetic Voice Data for Automatic Speech Recognition in African Languages](https://mlsanity.com/paper/2507.17578/) (2025)
- [When and How Human Curation Backfires: Preference Alignment under Multi-Model Self-Consuming Loop](https://mlsanity.com/paper/2605.29267/) (2026)
- [Knowledge Collapse in LLMs: When Fluency Survives but Facts Fail under Recursive Synthetic Training](https://mlsanity.com/paper/2509.04796/) (2025)
- [Position: the Stochastic Parrot in the Coal Mine. Model Collapse is a Threat to Low-Resource Communities](https://mlsanity.com/paper/2605.04127/) (2026)
- [Will Humanity Be Rendered Obsolete by AI?](https://mlsanity.com/paper/2510.22814/) (2025)
- [Equilibrium Dynamics and Mitigation of Gender Bias in Synthetically Generated Data](https://mlsanity.com/paper/2511.10689/) (2025)
- [The Rise of AI in Weather and Climate Information and its Impact on Global Inequality](https://mlsanity.com/paper/2603.05710/) (2026)
- [Iterative Finetuning is Mostly Idempotent](https://mlsanity.com/paper/2605.01130/) (2026)
- [To do(x) or not to do(x): Medical Image Counterfactuals for Dataset Augmentation](https://mlsanity.com/paper/2609.14124/) (2026)

## Abstract (arXiv)

Model collapse, a phenomenon characterized by performance degradation due to iterative training on synthetic data, has been widely studied. However, its implications for bias amplification, the progressive intensification of pre-existing societal biases in Large Language Models (LLMs), remain significantly underexplored, despite the growing influence of LLMs in shaping online discourse. In this paper, we introduce a open, generational, and long-context benchmark specifically designed to measure political bias amplification in LLMs, leveraging sentence continuation tasks derived from a comprehensive dataset of U.S. political news. Our empirical study using GPT-2 reveals consistent and substantial political bias intensification (e.g., right-leaning amplification) over iterative synthetic training cycles. We evaluate three mitigation strategies, Overfitting, Preservation, and Accumulation, and demonstrate that bias amplification persists independently of model collapse, even when the latter is effectively controlled. Furthermore, we propose a mechanistic analysis approach that identifies neurons correlated with specific phenomena during inference through regression and statistical tests. This analysis uncovers largely distinct neuron populations driving bias amplification and model collapse, underscoring fundamentally different underlying mechanisms. Finally, we supplement our empirical findings with theoretical intuition that explains the separate origins of these phenomena, guiding targeted strategies for bias mitigation.