arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2604.22503cs.CL
收藏

衡量和缓解AI写作辅助工具中的人格扭曲

How AI Writing Assistance Distorts Perceptions of AI Users and Their Opinions

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

Paul Röttger, Kobi Hackenburg, Hannah Rose Kirk, Christopher Summerfield

AI总结:

研究评估了AI写作辅助工具如何扭曲写作者的人格特征,通过实验发现AI使写作者显得更具观点性、能力更强和积极,但写作者对这些扭曲表示反对却仍偏好使用AI辅助文本,通过训练奖励模型缓解了部分扭曲,但影响了用户接受度。

摘要:

数百万人使用人工智能(AI)进行写作辅助。本文评估了AI写作辅助工具如何扭曲写作者的人格——他们的感知信念、个性和身份。在三个大规模实验中,2939名写作者在有无AI辅助的情况下撰写了政治意见段落。11091名读者在29个社会显著维度上盲评这些段落,涵盖政治意见、写作质量、写作者个性、情感和人口统计数据。AI写作辅助在所有维度上产生了人格扭曲:使用AI时,写作者显得更具观点性、能力更强和积极,其感知人口统计数据向更受特权群体倾斜。写作者反对许多观察到的扭曲,但即使意识到这些扭曲,仍偏好使用AI辅助文本。我们通过在实验数据(10008段落,2903596个评分)上训练奖励模型,在模型层面成功缓解了令人反感的人格扭曲,但此举影响了用户接受度,表明AI写作辅助的有益和有害属性之间存在复杂关系,可能难以解决。共同发现表明,即使在人类监督的现实条件下,AI写作辅助的人格扭曲也是普遍且持续的,这在AI普及扩大的背景下对公共讨论、信任和民主辩论具有影响。

英文摘要:

Hundreds of millions of people use AI for writing assistance. We asked how AI distorts perceptions of AI users and their opinions. Writers (N=2,939) wrote political opinion paragraphs with and without AI assistance. Separate readers (N=11,091) blindly evaluated these paragraphs and made inferences about the writers. With AI, writers seemed more opinionated, competent, positive, and demographically privileged. Writers objected to many of these distortions yet preferred AI-assisted text even when made aware of them. Reward models trained on our data mitigated objectionable distortions but reduced user acceptance. Additional readers (N=8,798) trusted AI-assisted writers more and were more persuaded by AI writing when AI was more distortive. Overall, distortions from AI writing assistance are pervasive, persistent under human oversight, and likely to have consequential effects on human behaviours and attitudes.

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。

↑