机器写作中的性别刻板印象:基于实验研究的实然探讨

Gender Stereotypes in Perceiving Machine-Written Content:Empirical Evidence Based on an Experimental Study

性别刻板印象会影响读者对作者与内容的评价。基于CASA(Computers are Social Actors)范式,本研究探讨在机器写作领域中,被赋予了性别的机器作者是否也会受到性别刻板印象的影响。一个采用因子设计的在线实验收集了读者对不同物种、不同性别作者及写作文本的评价。研究发现性别刻板印象的确存在于大众对机器作者及内容的评价中;但在具体类型的文本中,读者对文本质量和作者的评价呈现出物种和性别的差异。

Gender stereotypes can influence people’s assessments on writers and content quality.Based on the CASA paradigm,this study attempts to examine whether the AI-writers with different gender labels are affected by the gender stereotypes.An online experiment with factorial design was conducted to collect people’s evaluations on the content quality and expertise of the writers.The results indicated that gender stereotypes do exist in people’s assessments on AIwriters and content quality.But for specific contexts,readers’appraisals of the quality of the content and expertise of the writer vary based on the writers’species and genders.