MONAL: Model Autophagy Analysis for Modeling Human-AI Interactions

TL;DR

This paper introduces MONAL, a novel analysis method for understanding how large AI models process and potentially suppress human-generated information through self-consumption loops, revealing impacts on information diversity and model performance.

Contribution

MONAL provides a new framework for analyzing human-AI interactions and the self-consumption dynamics within large models, addressing a gap in understanding their societal and informational effects.

Findings

Synthetic information increasingly dominates training datasets over time.

Large models tend to selectively modify or prioritize certain contents.

Information diversity decreases, causing potential bottlenecks in model performance.

Abstract

The increasing significance of large models and their multi-modal variants in societal information processing has ignited debates on social safety and ethics. However, there exists a paucity of comprehensive analysis for: (i) the interactions between human and artificial intelligence systems, and (ii) understanding and addressing the associated limitations. To bridge this gap, we propose Model Autophagy Analysis (MONAL) for large models' self-consumption explanation. MONAL employs two distinct autophagous loops (referred to as ``self-consumption loops'') to elucidate the suppression of human-generated information in the exchange between human and AI systems. Through comprehensive experiments on diverse datasets, we evaluate the capacities of generated models as both creators and disseminators of information. Our key findings reveal (i) A progressive prevalence of model-generated synthetic information over time within training datasets compared to human-generated information; (ii) The discernible tendency of large models, when acting as information transmitters across multiple iterations, to selectively modify or prioritize specific contents; and (iii) The potential for a reduction in the diversity of socially or human-generated information, leading to bottlenecks in the performance enhancement of large models and confining them to local optima.