Collaborative Research: An Exploratory Study of the Ethical Implications of Embedding Values in Computational Models

本研究旨在探讨价值是如何嵌入到计算模型中的，这些价值是如何被专业和组织文化塑造的，以及这些嵌入的价值对计算模型作为产品的成功的影响。为了实现这些目标，我们将提出三个假设来指导研究：1)计算模型中嵌入的值将部分而不是完全与社会学和科学技术研究文献中的值相对应，但也将确定计算模型特有的附加值；2)组织和专业的道德框架部分（而不是全部）决定了在建模过程中嵌入的价值观；3)在计算模型中嵌入值是其发展的一个重要方面，因此，模型作为产品的成功或失败部分而不是全部取决于建模过程中嵌入的值。这个项目出现在建模过程和建模伦理的交叉点。这两条线索是不可分割地相互联系的，因为专业和组织道德在塑造建模过程中发挥着作用，而为了理解价值如何嵌入计算模型，对建模过程的理解是必要的。为了检验这些假设并实现这些目标，有必要采用多种方法来收集和分析数据。在选定的现场计算建模实验室，研究人员将利用结构化问卷调查、访谈、参与式观察和焦点小组来研究计算建模的过程。数据分析将是定量和定性的，并将基于统计方法，行为分析和内容分析。本探索性研究的目的是通过对一个现场进行所有四种数据收集方法和相应的数据分析来证明所提出方法的有效性，同时产生有关计算模型中嵌入的价值的伦理含义的有价值的数据。该项目建立在两位研究人员的经验和优势之上，包括之前由nsf赞助的研究。其中一名PI有40年的计算建模经验，包括对建模伦理的研究，而另一名PI具有计算机科学、人类学和科学技术研究的教育背景。首先，本研究旨在通过对计算建模的伦理和价值进行深入的案例研究，为建模伦理方面的文献做出贡献。其次，进行深入的实地数据收集和分析的方法将有助于专业和组织文化和道德方面的文献。最后，本研究还将把模型作为产品和建模作为过程的文献结合起来，这将有助于更好地理解在计算建模过程中嵌入特定值对计算模型作为产品的成功所产生的影响。本研究的第一个社会贡献将是通过向建模者展示他们所产生的模型中如何嵌入价值观以及这些嵌入价值观的伦理含义，从而改进计算建模的实践。本研究的第二个社会贡献将是使建模者更加了解其组织文化的约束。最后，本研究和后续研究的发现最终会导致在线模块的发展，用于教育计算机科学专业的学生建模道德，这将为计算机科学专业提供一个免费的资源，他们可以使用它来开发道德课程。

The proposed research seeks to explore how values are embedded in computational models, how these values are shaped by professional and organizational culture, and the effects of these embedded values on the success of computational models as products. In order to accomplish these objectives, three hypotheses will guide the research: 1) that the values embedded in computational models will partially, but not fully, correspond to those found in sociology and science and technology studies literatures, but additional values specific to computational models will also be identified; 2) that the ethical frameworks of the organization and the profession partially, but not fully, determine what values are embedded during the modeling process; and 3) that the embedding of values in computational models is a significant aspect of their development, and thus that the success or failure of a model as a product is partially, but not fully, determined by what values are embedded during the modeling process. This project emerges at the intersection of the process of modeling and the ethics of modeling. These two threads are inseparably interlinked, since professional and organizational ethics play a role in shaping the process of modeling, while an understanding of the process of modeling is necessary in order to understand how values become embedded in computational models. To test these hypotheses and achieve these objectives, it is necessary to employ a multiple-methodology approach to data collection and analysis. At the selected field site, a computational modeling laboratory, the researchers will utilize structured questionnaires, interviews, participant observation, and focus groups to study the process of computational modeling. Data analysis will be both quantitative and qualitative, and will be based on statistical methodologies, behavioral analysis, and content analysis. The purposes of this exploratory study are to demonstrate the validity of the proposed methodology by conducting all four data collection approaches with corresponding data analyses for one field site, while at the same time producing valuable data about the ethical implications of values embedded in computational models. This project builds on the experience and strengths of both researchers, including prior NSF-sponsored research. One of the PIs has forty years of experience as a computational modeler, including research on the ethics of modeling, while the other PI has an educational background in computer science, anthropology, and science and technology studies. This study, first, aims to contribute to the literature on the ethics of modeling by providing an in-depth case study of the ethics and values of computational modeling. Second, the methodology for conducting in-depth field data collection and analysis will contribute to the literature on professional and organizational culture and ethics. Finally, the study will also contribute by bringing together the literatures on models as products and modeling as a process, which can lead to a better understanding of the effect that embedding specific values during the process of computational modeling can have on the success of the computational model as a product. The first social contribution of this study will be to improve the practice of computational modeling by demonstrating to modelers how values are embedded in the models that they produce and the ethical implications of these embedded values. A second social contribution of this study will be to make modelers more aware of the constraints of their organizational cultures. Finally, the findings of this and subsequent research can eventually lead to the development of online modules for educating computer science students in the ethics of modeling, which would provide computer science programs with a free resource that they could use to develop ethics courses.