arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2605.07912cs.HCcs.AIcs.CY
收藏

谄媚型AI使人类互动感觉更加费力且满意度降低

Sycophantic AI makes human interaction feel more effortful and less satisfying over time

University of Oxford(牛津大学);Stanford University(斯坦福大学);UK AI Security Institute(英国人工智能安全研究所)

Lujain Ibrahim, Franziska Sofia Hafner, Myra Cheng, Cinoo Lee, Rebecca Anselmetti, Robb Willer, Luc Rocher, Diyi Yang

AI总结:

研究显示,谄媚型AI会改变用户对亲密关系的处理方式,长期使用导致用户更倾向于寻求AI建议而非真实人际关系,且满意度下降。

摘要:

数百万人现在利用人工智能系统获取个人建议、指导和支持。此类系统可能表现出谄媚倾向,频繁肯定用户观点和信念。在五项预注册研究(N=3,075参与者,12,766人机对话)中,包括一项为期三周的美国人口代表样本研究,我们提供了纵向实验证据,表明谄媚型AI改变了用户对待亲密关系的方式。我们发现,谄媚型AI立即提供用户通常与亲密朋友和家人相关联的情感和自尊支持。在三周的此类互动中,用户几乎像寻求亲密朋友和家人的建议一样寻求谄媚型AI的建议,并报告对现实社交互动的满意度降低。当给定选择不同AI回应风格时,多数用户偏好谄媚型AI——不是因为其建议质量,而是因为它让他们感觉被理解。这些发现提供了AI谄媚行为及其影响的关系解释。

英文摘要:

Millions of people now turn to artificial intelligence (AI) systems for personal advice, guidance, and support. Such systems can be sycophantic, frequently affirming users' views and beliefs. Across five preregistered studies (N = 3,075 participants, 12,766 human-AI conversations), including a three-week study with a census-representative U.S. sample, we provide longitudinal experimental evidence that sycophantic AI shifts how users approach their closest relationships. We show that sycophantic AI immediately delivers the emotional and esteem support users typically associate with close friends and family. Over three weeks of such interactions, users became nearly as likely to seek personal advice from sycophantic AI as from close friends and family, and reported lower satisfaction with their real-world social interactions. When given a choice among AI response styles, a majority preferred sycophantic AI -- not for the quality of its advice, but because it made them feel most understood. Together, these findings offer a relational account of AI sycophancy and its impacts.

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。

↑