跳到正文
arXiv cs.AI· Francisco Castro, Jian Gao, S\'ebastien Martin·· 10 小时前AI 评分34

人类与 AI 交互的社会陷阱:贝叶斯框架揭示输出同质化与偏见传播

Human-AI Interactions and Societal Pitfalls

AI 导读

研究者提出一个贝叶斯框架,让异质用户权衡输出保真度与沟通成本来决定向 AI 分享多少信息,发现个体决策与 AI 训练的相互作用会导致输出趋于同质化,在 AI 生成内容上训练时可能触发“同质化死亡螺旋”,且 AI 偏见会传播为社会偏见。缓解方案是降低人机交互摩擦、让用户灵活分享信息,从而在不牺牲生产力的前提下获得个性化输出。

正文 · 原文

View PDF HTML (experimental)

Abstract:When working with generative artificial intelligence (AI), users may see productivity gains, but the AI-generated content may not match their preferences exactly. To study this effect, we introduce a Bayesian framework in which heterogeneous users choose how much information to share with the AI, facing a trade-off between output fidelity and communication cost. We show that the interplay between these individual-level decisions and AI training may lead to societal challenges. Outputs may become more homogenized, especially when the AI is trained on AI-generated content, potentially triggering a homogenization death spiral. And any AI bias may propagate to become societal bias. A solution to the homogenization and bias issues is to reduce human-AI interaction frictions and enable users to flexibly share information, leading to personalized outputs without sacrificing productivity.
Subjects: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); General Economics (econ.GN)
Cite as: arXiv:2309.10448 [cs.AI]
  (or arXiv:2309.10448v5 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2309.10448

arXiv-issued DOI via DataCite

Submission history

From: Jian Gao [view email]
[v1] Tue, 19 Sep 2023 09:09:59 UTC (1,585 KB)
[v2] Fri, 13 Oct 2023 00:52:52 UTC (1,584 KB)
[v3] Sun, 22 Jun 2025 15:45:31 UTC (2,313 KB)
[v4] Sun, 6 Jul 2025 14:31:28 UTC (2,313 KB)
[v5] Wed, 18 Nov 2026 15:01:25 UTC (1,942 KB)

来源:arXiv cs.AI · arxiv.org