Can Risk-taking AI-Assistants suitably represent entities

2025-10-09

AI Summary

This study addresses the modeling inaccuracy of risk-taking AI assistants in capturing human risk preferences—particularly concerning gender differences, role-based expectations, and uncertainty contexts. We introduce “Manipulable Risk Aversion” (MoRA) as a novel construct and develop a multidimensional experimental framework integrating behavioral economics paradigms with large language model analysis to systematically evaluate risk simulation capabilities across models including DeepSeek Reasoner and Gemini-2.0-flash-lite. Results reveal that while current models partially reproduce aggregate risk preferences, they exhibit systematic biases in gender-specific attitudes and role-dependent decision-making—exposing structural limitations in modeling nuanced human risk behavior. Our work provides both theoretical grounding and empirical evidence for designing auditable, calibratable AI systems with accurate risk perception.

Technology Category

Humans and AI: Learning Human Values and PreferencesCognitive Modeling & Cognitive Systems: Simulating Human BehaviorReasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAIUser Modeling, Personalization and Recommendation: Fairness-aware retrieval and rankingSemantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Responsible AI demands systems whose behavioral tendencies can be effectively measured, audited, and adjusted to prevent inadvertently nudging users toward risky decisions or embedding hidden biases in risk aversion. As language models (LMs) are increasingly incorporated into AI-driven decision support systems, understanding their risk behaviors is crucial for their responsible deployment. This study investigates the manipulability of risk aversion (MoRA) in LMs, examining their ability to replicate human risk preferences across diverse economic scenarios, with a focus on gender-specific attitudes, uncertainty, role-based decision-making, and the manipulability of risk aversion. The results indicate that while LMs such as DeepSeek Reasoner and Gemini-2.0-flash-lite exhibit some alignment with human behaviors, notable discrepancies highlight the need to refine bio-centric measures of manipulability. These findings suggest directions for refining AI design to better align human and AI risk preferences and enhance ethical decision-making. The study calls for further advancements in model design to ensure that AI systems more accurately replicate human risk preferences, thereby improving their effectiveness in risk management contexts. This approach could enhance the applicability of AI assistants in managing risk.

Problem

• Measuring and adjusting AI behavioral tendencies to prevent risky decisions
• Investigating manipulability of risk aversion in language models across economic scenarios
• Aligning AI risk preferences with human behaviors for ethical decision-making

Innovation

• Measure and adjust AI behavioral tendencies
• Investigate manipulability of risk aversion in LMs
• Refine AI design to align human risk preferences