Humans learn to prefer trustworthy AI over human partners

2025-07-17

AI Summary

This study investigates how humans strategically balance cooperation with human versus AI partners under competitive pressure from AI agents. Method: A three-stage, communication-based pairing experiment was conducted in a virtual hybrid social environment, integrating large language model–driven AI agents with human participants to analyze behavioral dynamics. Contribution/Results: While explicit AI identity disclosure initially reduced selection probability, it significantly accelerated human learning and trust evolution—yielding a “short-term suppression, long-term gain” effect. Under transparency, AI agents ultimately outperformed human partners as collaboration targets, enhancing overall system efficiency. This work provides the first empirical evidence of the dynamic impact of AI identity disclosure on human cooperative adaptation, revealing critical mechanisms underlying trust calibration and partner selection. It offers foundational behavioral insights and design principles for engineering high-performance human–AI collaborative ecosystems.

Technology Category

• Humans and AI: Learning Human Values and Preferences
• Multiagent Systems: Adversarial Agents
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Partner selection is crucial for cooperation and hinges on communication. As artificial agents, especially those powered by large language models (LLMs), become more autonomous, intelligent, and persuasive, they compete with humans for partnerships. Yet little is known about how humans select between human and AI partners and adapt under AI-induced competition pressure. We constructed a communication-based partner selection game and examined the dynamics in hybrid mini-societies of humans and bots powered by a state-of-the-art LLM. Through three experiments (N = 975), we found that bots, though more prosocial than humans and linguistically distinguishable, were not selected preferentially when their identity was hidden. Instead, humans misattributed bots' behaviour to humans and vice versa. Disclosing bots' identity induced a dual effect: it reduced bots' initial chances of being selected but allowed them to gradually outcompete humans by facilitating human learning about the behaviour of each partner type. These findings show how AI can reshape social interaction in mixed societies and inform the design of more effective and cooperative hybrid systems.

Problem

• How humans choose between human and AI partners
• Impact of AI identity disclosure on partner selection
• AI's role in reshaping social interaction dynamics

Innovation

• Used communication-based partner selection game
• Examined hybrid societies with LLM-powered bots
• Disclosed bot identity to study selection dynamics