The ethics of scientific recommender systems
- PDF / 533,315 Bytes
- 8 Pages / 439.37 x 666.142 pts Page_size
- 1 Downloads / 189 Views
The ethics of scientific recommender systems Andrea Polonioli1 Received: 3 July 2020 / Accepted: 14 October 2020 © Akadémiai Kiadó, Budapest, Hungary 2020
Abstract Scientific recommender systems have become increasingly popular as a tool to overcome information overload, allowing researchers to access fresh and relevant content. However, this article presents an analysis of the most pressing ethical challenges posed by recommender systems in the context of scientific research. In particular, it is argued that scientific recommender systems may risk isolating scholars in information bubbles and insulating them from exposure to different viewpoints. Further, they also risk suffering from popularity biases which may lead to a winner-takes-all scenario and reinforce discrepancies in recognition received by eminent scientists and unknown researchers. The article concludes with recommendations for scientists, journals, and digital libraries to facilitate progress in the study of the ethics of scientific recommender systems. Recent technological advancements have altered many aspects of life significantly. Tasks previously left to humans are now increasingly delegated to algorithms. In particular, profiling algorithms have recently garnered a great deal of attention in areas such as commerce, media and entertainment, where people are provided with suggestions about what to buy, what to read or where to go (Mittelstadt et al. 2016). For instance, Amazon, Netflix as well as news aggregators provide nice illustrations of recommender systems we frequently use and interact with. One particular context in which recommender systems have become increasingly popular is scientific research (Sugiyama and Kan 2010; Beel et al. 2016). For instance, leading digital libraries such as Elsevier, PubMed, SpringerLink all have systems that can send out email alerts or provide feeds on paper recommendations that match user interests, but many others are actually available and are being used by scientists across different fields (Gibney 2014). While different names are used in the literature, we will adopt the terminology of scientific recommender systems (SRS) to refer to those filtering systems that take information about a user’s preferences, profile and behaviour as an input, and output a prediction about the rating that a user would give of the scholarly content under evaluation, and predict how they would rank a set of scholarly items individually or as a bundle.
* Andrea Polonioli [email protected] 1
AI Labs, Coveo Solutions Inc, 3175 des Quatre‑Bourgeois Suite 200, Quebec City, QC G1W 2K7, Canada
13
Vol.:(0123456789)
Scientometrics
This paper introduces a new research area that lies at the intersection of information ethics, information and communication technology and the ethics of scientific research. Specifically, we introduce the ethics of scientific recommendation systems as a highly important topic of investigation and encourage more work and research in this area. The rationale for introducing the topic i
Data Loading...
