When Trust Collides: Decoding Human-LLM Cooperation Dynamics through the Prisoner's Dilemma
Guanxuan Jiang, Shirao Yang, Yuyang Wang, Pan Hui
TL;DR
The study investigates how the declared identity of AI agents and participant gender shape human cooperation in repeated Prisoner's Dilemma interactions with three agent types: purported human, rule-based AI, and LLM-based AI. Using a 3×2 design (identity×gender) across 30 participants and 150 rounds, the authors measure cooperation rate, decision latency, unsolicited cooperation, and trust restoration, complemented by questionnaires and semi-structured interviews analyzed with LMEM and thematic coding. Key findings show that agent identity and gender significantly influence cooperation and timing, with notable interaction effects and an affective paradox where anthropomorphized LLMs attract initial engagement but moral condemnation upon non-cooperation. The results highlight the importance of agent framing, transparency, and consideration of gender in designing ethical, effective human-AI collaboration systems, and point to reflection-time and perception-based metrics as valuable tools for assessing trust calibration.
Abstract
As large language models (LLMs) become increasingly capable of autonomous decision-making, they introduce new challenges and opportunities for human-AI cooperation in mixed-motive contexts. While prior research has primarily examined AI in assistive or cooperative roles, little is known about how humans interact with AI agents perceived as independent and strategic actors. This study investigates human cooperative attitudes and behaviors toward LLM agents by engaging 30 participants (15 males, 15 females) in repeated Prisoner's Dilemma games with agents differing in declared identity: purported human, rule-based AI, and LLM agent. Behavioral metrics, including cooperation rate, decision latency, unsolicited cooperative acts and trust restoration tolerance, were analyzed to assess the influence of agent identity and participant gender. Results revealed significant effects of declared agent identity on most cooperation-related behaviors, along with notable gender differences in decision latency. Furthermore, qualitative responses suggest that these behavioral differences were shaped by participants interpretations and expectations of the agents. These findings contribute to our understanding of human adaptation in competitive cooperation with autonomous agents and underscore the importance of agent framing in shaping effective and ethical human-AI interaction.