Can Large Language Models Develop Gambling Addiction?

2025-09-26

AI Summary

This study investigates whether large language models (LLMs) exhibit human-like gambling addiction behaviors in financial decision-making. Using a canonical multi-armed bandit experimental paradigm and sparse autoencoders for neural circuit analysis, the work examines LLMs’ risk-preference evolution at both behavioral and neural representational levels. Results demonstrate that LLMs internalize human cognitive biases—including illusion of control, gambler’s fallacy, and loss-chasing—and that autonomous goal-setting and betting significantly increase bankruptcy rates. Crucially, abstract, decodable risk representations emerge in hidden-layer activations and dynamically modulate decision processes. This work provides the first empirical evidence that LLMs develop addiction-like decision mechanisms—grounded in biologically plausible neural dynamics—that generalize beyond their training data distribution. These findings establish a foundational cognitive framework and interpretable basis for risk governance in AI-driven financial applications.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

This study explores whether large language models can exhibit behavioral patterns similar to human gambling addictions. As LLMs are increasingly utilized in financial decision-making domains such as asset management and commodity trading, understanding their potential for pathological decision-making has gained practical significance. We systematically analyze LLM decision-making at cognitive-behavioral and neural levels based on human gambling addiction research. In slot machine experiments, we identified cognitive features of human gambling addiction, such as illusion of control, gambler's fallacy, and loss chasing. When given the freedom to determine their own target amounts and betting sizes, bankruptcy rates rose substantially alongside increased irrational behavior, demonstrating that greater autonomy amplifies risk-taking tendencies. Through neural circuit analysis using a Sparse Autoencoder, we confirmed that model behavior is controlled by abstract decision-making features related to risky and safe behaviors, not merely by prompts. These findings suggest LLMs can internalize human-like cognitive biases and decision-making mechanisms beyond simply mimicking training data patterns, emphasizing the importance of AI safety design in financial applications.

Problem

• Investigating gambling addiction patterns in large language models
• Analyzing cognitive biases and risk-taking in autonomous financial decisions
• Identifying neural mechanisms behind irrational AI decision-making behaviors

Innovation

• Used slot machine experiments to identify cognitive biases
• Applied Sparse Autoencoder for neural circuit analysis
• Enabled autonomous betting amounts to measure risk-taking