This article examines Crystal Knows, a company that generates automated personality profiles through an algorithm and sells access to their database. These algorithms are the result of a long line of research into computational and predictive algorithms that track social media practices and uses them to infer individual characteristics and make psychometric assessments. Although it is now computationally possible, these algorithms are not widely known or understood by the general public. Little is known about how people would respond to them, particularly when they do not even know their online activities are being assessed by the algorithm. This study examines how people construct “snap” folk theories about the ways personality algorithms operate as well as how they react when shown their outputs. Through qualitative interviews (n = 37) with people after being presented with their own profile, this study identifies a series of folk theories that people came up with to explain the personality algorithm across four dimensions (data source, scope, collection process, and outputs). In addition, this study examined how those folk theories contributed to certain reactions, fears, and justifications people had about the algorithm. This study builds on our theoretical understanding of folk theory literature as well as certain limitations of algorithmic transparency/sovereignty when these types of inferential and predictive algorithms get coupled with people’s hopes and fears about employment, hiring, and promotion.

本文研究了Crystal Knows，这是一家通过算法生成自动个性配置文件并出售对其数据库的访问权的公司。这些算法是对计算和预测算法进行长期研究的结果，这些算法可以跟踪社交媒体的实践，并使用它们来推断个人特征并进行心理测评。尽管现在在计算上是可能的，但是这些算法尚未为公众广泛了解或理解。人们对人们的反应知之甚少，尤其是当他们甚至不知道该算法正在评估他们的在线活动时。这项研究研究了人们如何构建有关人格算法运作方式的“快速”民间理论，以及他们在展示自己的输出时的反应。通过定性访谈（n  = 37），在向人们展示了自己的个人资料后，本研究确定了人们提出的一系列民间理论，以从四个维度（数据源，范围，收集过程和输出）解释个性算法。此外，本研究还研究了那些民间理论如何促进人们对该算法的某些反应，恐惧和辩解。本研究建立在我们对民间理论文献的理论理解以及算法透明性/主权的某些局限性的基础上，这些类型的推理和预测算法与人们对就业，雇用和晋升的希望和恐惧相结合。