arXiv Machine Learning June 2, 2026 By Devon Jarvis, Richard Klein, Benjamin Rosman, Steven James, Stefano Sarao Mannelli

Position: the Stochastic Parrot in the Coal Mine. Model Collapse is a Threat to Low-Resource Communities

Read the original on arXiv Machine Learning →

arXiv:2605. 04127v2 Announce Type: replace Abstract: Model collapse, the degradation in performance that arises when generative models are trained on the outputs of prior models, is an increasing concern as artificially generated content proliferates.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv Machine Learning.

llms

arXiv Machine Learning

Jul 31

Epistemic diversity across language models mitigates knowledge collapse

arXiv:2512. 15011v3 Announce Type: replace Abstract: Artificial intelligence (AI) increasingly generates the very content used to train future AI systems.

By Damian Hodel, Jevin D. West

llms

arXiv AI

Jul 1

Learning by Surprise: Adaptive Mitigation of Model Collapse in Large Language Models

arXiv:2410. 12341v4 Announce Type: replace-cross Abstract: As AI-generated content increasingly populates the web, generative AI models are at growing risk of being trained on their own outputs, a process known as AI autophagy.

By Daniele Gambetta, Gizem Gezici, Fosca Giannotti, Dino Pedreschi, Alistair Knott, Luca Pappalardo

llms fine-tuning

arXiv AI

Jun 6

Geographic Bias and Diversity in AI Evaluation

arXiv:2606. 05187v1 Announce Type: cross Abstract: Among the many challenges hindering the responsible development and deployment of AI, arguably none has faced more intense scrutiny than bias in its various forms.

By Zilong Liu, Krzysztof Janowicz, Gengchen Mai, Song Gao, Rui Zhu

llms benchmarks safety

arXiv AI

Jun 29

Can Generative Artificial Intelligence Survive Data Contamination? Theoretical Guarantees under Contaminated Recursive Training

arXiv:2602. 16065v2 Announce Type: replace-cross Abstract: As artificial intelligence (AI)-generated content proliferates, models are increasingly trained on their own outputs, risking progressive degradation or collapse.

By Kevin Wang, Hongqian Niu, Didong Li

safety

arXiv AI

Jul 7

Large Language Models Develop Novel Social Biases Through Adaptive Exploration

arXiv:2511. 06148v4 Announce Type: replace-cross Abstract: As large language models (LLMs) are adopted into frameworks that grant them the capacity to make real decisions, it is increasingly important to ensure that they are unbiased.

By Addison J. Wu, Ryan Liu, Xuechunzi Bai, Thomas L. Griffiths

llms safety

arXiv Machine Learning

Jun 2

Position: Stop Preaching and Start Practising Data Frugality for Responsible Development of AI

arXiv:2602. 19789v2 Announce Type: replace Abstract: This position paper argues that the machine learning community must move from preaching to practising data frugality for responsible artificial intelligence (AI) development.

By Sophia N. Wilson, Andrew Millard, Gu{\dh}r\'un Fj\'ola Gu{\dh}mundsd\'ottir, Raghavendra Selvan, Sebastian Mair

safety

Related stories

Epistemic diversity across language models mitigates knowledge collapse

Learning by Surprise: Adaptive Mitigation of Model Collapse in Large Language Models

Geographic Bias and Diversity in AI Evaluation

Can Generative Artificial Intelligence Survive Data Contamination? Theoretical Guarantees under Contaminated Recursive Training

Large Language Models Develop Novel Social Biases Through Adaptive Exploration

Position: Stop Preaching and Start Practising Data Frugality for Responsible Development of AI