TL;DR
This paper argues that information science, with its human-centered and transdisciplinary approach, offers crucial ethical and practical insights for the burgeoning field of data science. By reviewing existing literature, it highlights how information science can help address critical issues like data bias and promote a more humanistic approach to data science research and practice.
What Problem Does This Paper Address?
The rapid growth of data science has brought immense attention to its capabilities, but also to significant ethical challenges, particularly concerning data bias. While data science is rapidly developing, there's a perceived gap in understanding how established disciplines, like information science, can contribute to addressing these ethical concerns. Information science, with its long history of focusing on the human aspects of information, its lifecycle, and ethical considerations, is uniquely positioned to offer valuable perspectives and methodologies to data science, which often leans towards a more technical and quantitative approach. This paper aims to bridge this gap by reviewing the literature and demonstrating the contributions information science can make to data science ethics.
Methods
This paper is a narrative review of existing literature. The authors systematically searched three databases (Library and Information Science Source, Computer Source, and Google Scholar) for relevant articles published between July 2021 and April 2022. The selection of articles was guided by two central questions: 1) How do data bias and the evolution of data ethics appear in information science literature? and 2) What perspectives on data science bias and ethics within information science appear in the literature? The review focuses on synthesizing key findings and concepts rather than providing an exhaustive scoping review.
Datasets
Not discussed in the paper. This paper is a literature review and does not present new empirical research or datasets.
Key Insights and Ideas
- Information Science as a Humanistic Counterpoint: Information science, with its roots in librarianship and a focus on human-centered perspectives, offers a vital counterbalance to the often purely technical and quantitative focus of data science.
- Data Ethics is Not New to Information Science: The discipline has a long history of grappling with ethical considerations related to information, including privacy, bias, and intellectual freedom, providing a rich foundation for addressing data science ethics.
- The Data Life Cycle and Bias: Understanding the various stages of the data life cycle (generation, collection, processing, storage, analysis, visualization, interpretation) is crucial for identifying and mitigating different types of data bias (sampling, availability, methodological, contextual, interpretation).
- Information Science's Unique Contributions: Information scientists bring expertise in data curation, knowledge organization, and understanding information behavior, which are essential for a holistic approach to data science that goes beyond mere technical implementation.
- The "Humanistic Data Science" Movement: There's a growing trend within information science towards advocating for a more humanistic approach to data science, emphasizing the impact of data on people and society.
Key Results
The review highlights that information science researchers have already made significant contributions to the discourse on data bias and ethics. Key areas of contribution include:
- Ethical Frameworks: Information science has a history of developing and adhering to ethical codes that can be adapted to data science.
- Understanding Data Bias: The literature within information science offers detailed analyses of various types of data bias and their origins within the data life cycle.
- Human-Centered Perspectives: Information science emphasizes the importance of context, user needs, and the societal impact of data, which are crucial for developing responsible data science practices.
- Data Curation and Management: Information science expertise in managing information throughout its lifecycle is directly applicable to the challenges of data management and curation in data science.
Limitations and Future Work
The authors acknowledge that the interest in data science ethics within information science is still growing, and while a substantial body of work exists, it may not be exhaustive. They suggest that future work could delve deeper into the unique contributions of information science research and researchers to anti-bias and ethics literature in data science. The paper also implicitly points to the need for continued integration and collaboration between information science and data science fields to ensure ethical development and application of data-driven technologies.
Summary updated