arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2511.06148cs.CYcs.AIcs.CL
收藏

大语言模型通过适应性探索形成新的社会偏见

Large Language Models Develop Novel Social Biases Through Adaptive Exploration

Princeton University(普林斯顿大学)

Addison J. Wu, Ryan Liu, Xuechunzi Bai, Thomas L. Griffiths

AI总结:

研究大语言模型在无固有差异时能否形成新社会偏见,用心理学范式证明其能形成,导致任务分配不公平,新模型更严重,还探索干预措施,发现激励探索可减少分层。

摘要:

随着大语言模型(LLMs)被应用于能做出实际决策的框架中,确保其无偏见变得越发重要。本文认为仅从模型中消除现有偏见的主流方法是不够的。我们用心理学文献中的范式证明,即使不存在固有差异,LLMs也能自发形成关于人工人口群体的新社会偏见。这些偏见导致高度分层的任务分配,比人类参与者的分配更不公平,且在更新、更大的模型中更严重。在人类中,此类新出现的偏见已被证明是由探索-利用权衡导致的,即决策者探索太少,使早期观察强烈影响对整个人口群体的印象。为减轻这种影响,我们探索了一系列针对模型输入、问题结构和明确引导的干预措施。虽然大多数干预措施效果有限,但明确激励探索能有力减少分层,凸显了需要更好的多方面目标来减轻偏见。这些结果表明,LLMs不仅是人类社会偏见的被动反映,还能从经验中积极创造新的偏见,引发了关于这些系统将如何随时间塑造社会的紧迫问题。

英文摘要:

As large language models (LLMs) are adopted into frameworks that grant them the capacity to make real decisions, it is increasingly important to ensure that they are unbiased. In this paper, we argue that the predominant approach of simply removing existing biases from models is not enough. Using a paradigm from the psychology literature, we demonstrate that LLMs can spontaneously develop novel social biases about artificial demographic groups even when no inherent differences exist. These biases result in highly stratified task allocations, which are less fair than assignments by human participants and are exacerbated in newer and larger models. In humans, emergent biases like these have been shown to result from exploration-exploitation trade-offs, where the decision-maker explores too little, allowing early observations to strongly influence impressions about entire demographic groups. To alleviate this effect, we explore a series of interventions targeting model inputs, problem structure, and explicit steering. While most interventions have limited effect, explicitly incentivizing exploration robustly reduces stratification, highlighting the need for better multifaceted objectives to mitigate bias. These results reveal that LLMs are not merely passive mirrors of human social biases, but can actively create new ones from experience, raising urgent questions about how these systems will shape societies over time.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑