Elsevier

Journal of Engineering and Technology Management

Volume 50, October–December 2018, Pages 45-60
Journal of Engineering and Technology Management

Proactive development of emerging technology in a socially responsible manner: Data-driven problem solving process using latent semantic analysis

https://doi.org/10.1016/j.jengtecman.2018.10.001Get rights and content

Abstract

This article offers a quantitative method of realizing the responsible development of emerging technologies via text analysis of future-oriented web data and scientific publication data. Latent Semantic Analysis (LSA) is applied to group key terms based on their contextual similarity. It overlooks what people have in mind regarding social disruptions of new technologies and further provides technical directions of how pertinent human values could be translated into their design in order to prevent the ramifications from occurring. Such a clear-cut methodology may allow the developers of emerging technologies to consider social concerns and human norms alongside more typical engineering ideals.

Introduction

Considering the great history of technology, no person can deny that new and emerging technologies have resolved numerous challenges of humanity. Meanwhile, some minds think from a different point of view; they address the destructive nature of those technologies, including global warming, privacy, and security hazards. These unprecedented consequences are reconstructing our existing societal, economic, and environmental frames and transforming the fundamental conditions of our lives and moral assumptions (Marx, 2010). As such, technology is holding a paradoxical nature: a novel solution for societal challenges but, at the same time, a critical agent that could cause far-reaching chaos. In response to these ramifications, we say the notion of responsibility must be incorporated at the earliest phase of technology’s innovation process. The responsibility points out our neglect of technology’s unpredictable consequences on individuals, societies, and environments in favor of immediate commercial success and economic growth (Grunwald, 2011; Blok and Lemmens, 2015). In the field of responsible innovation (RI), scholars have asserted that technology must be shaped or designed according to certain social values, in the hope that unintended and often undesirable side effects would never arise in the first place. However, one criticism of much of the literature on responsible development is that there exists a tendency of separating thoughts from actions (Firat et al., 2008; Grunwald and Achternbosch, 2013). The field is still at its embryonic stage, and the research to date has focused on vigorous clarification of responsible innovation rather than the establishment of specific guidelines that could actually be utilized in real life.

The notion of acting on responsibility in technology development has spurred few initiatives, such as Value Sensitive Design (VSD) and Value Conscious Design (Flanagan and Nissenbaum, 2014). Through gathering and analyzing information from field experts through participatory approaches, conventional studies have offered systematic problem-solving processes and have led to a renewed interest in the fields of Technology Assessment (TA) and Technology Foresight (TF). However, there still exist several drawbacks of the conventional participatory approaches: inefficiency and ineffectiveness. It is labor intensive and time-consuming to assemble experts from diverse fields and directly collect related data for the study. The results are also usually derived in a foreseeable manner with a very limited amount of invaluable information. This research, in response, attempts to approach in a more quantitative fashion by incorporating two data sources: future-oriented web data and literature data. Latent Semantic Analysis (LSA) text mining technique is applied to extract insightful topics or concepts based on the terms’ semantic relations. Generated by Web 2.0, future-oriented web information encapsulates the opinions of futures experts and the public interested in future technology and its societal change. Produced based on the visions of eminent scholars from a range of disciplines, scientific publication data yields novel solutions associated with the derived issues of emerging technologies.

The proposed method envisions a set of eventualities where a newborn technology plays out to be a harmful tool to the society and suggests essential technical requirements that must be focused in the design process for the right technological preparedness. Based on this research, numerous unforeseen impacts could be prevented ex-ante and thereby promoting only the right impact (von Schomberg, 2013; Owen et al., 2012), or the right effect, to our society.

Section snippets

Theoretical paradigm shifts towards responsible development

Quite different from established technologies, emerging technologies alter our existing social norms and values. Their distinctive characteristics have gradually affecting and thus changing how people perceive and react to the consequential nature of new technologies.

Participatory approach

New and emerging technology is characterized by its inherently distinctive natures: uncertainty and complexity (Maine et al., 2014). Adoption rate at the early stage of technology’s lifecycle or promised return from the investment is unclear (Halaweh, 2013). In addition, it is challenging to fully understand cause and effect chains and network effects among the actors associated with new technologies (Köhler and Som, 2014). This social uncertainty is the key component that causes such

Concept extraction using LSA

How could we then manage to better understand a massive collection of texts and point out only the key concepts or topics? Various methodologies can be accommodated in response, including Latent Semantic Analysis (LSA), Probabilistic Latent Semantic Analysis (pLSA), Latent Dirichlet Allocation (LDA), and Principal Components Analysis (PCA) (De Melo and Siersdorfer, 2007; Lee et al., 2010; Newman et al., 2014). These approaches measure co-occurrence relationships among terms and extract

Research framework

The overview of the research framework is illustrated in Fig. 2. The first step is envisioning social problems based on future-oriented web data, and the second step is deriving technical solutions for each social problem based on scientific publication data. The following sections will provide detailed explanations of sub-steps.

Illustrative case example: autonomous vehicle

In order to illustrate the applicability and effectiveness of our proposed approach, we have conducted a case study with an autonomous vehicle (AV) technology. AV was chosen as the target technology since its intrinsic nature corresponded well with our aforementioned definition of emerging technology. First, while AV has enormous potential to create social and economic opportunities in terms of road safety, urban development, energy consumption, or driver productivity (Greenblatt and Saxena,

Generating impact scenarios

The text analysis of future-oriented web data not only reduced time and effort in overviewing the concerns of emerging technologies but also led to a whole new approach of developing future scenarios. A scenario is a storyline, or a short novel, which illustrates a series of future eventualities. It is normally composed of one or two paragraphs; each paragraph involves multiple sentences; and each sentence comprises a series of terms and phrases. However, in order to construct one with

Concluding remarks

When it comes to the responsible development of emerging technologies, there has been a prime gap between philosophizing and existing. A guidance or a clear-cut methodology was in need for engineers to develop them with some sort of responsibility. The developers clearly knew the technology will ultimately transform the society, but couldn’t grasp what values are specifically affected and how they could be embedded into the development process to prevent such radical impacts from occurring. As

Acknowledgement

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (NO. 2011-0030814).

References (65)

  • F. Vanclay

    Conceptualising social impacts

    Environ. Impact Assess. Rev.

    (2002)
  • H. Xu et al.

    A value sensitive design investigation of privacy enhancing tools in web browsers

    Decis. Support Syst.

    (2012)
  • M. Zhu et al.

    Automatic dimensionality selection from the scree plot via the use of profile likelihood

    Comput. Stat. Data Anal.

    (2006)
  • C. Andronis et al.

    Literature mining, ontologies and information visualization for drug repurposing

    Brief. Bioinf.

    (2011)
  • R. Blake et al.

    Analyzing information systems security research to find key topics, trends, and opportunities

    J. Inf. Priv. Secur.

    (2012)
  • V. Blok et al.

    The emerging concept of responsible innovation. Three reasons why it is questionable and calls for a radical transformation of the concept of innovation

  • R.B. Bradford

    An empirical study of required dimensionality for large-scale latent semantic indexing applications

    ACM Conference on Information and Knowledge Management

    (2008)
  • Y. Chen et al.

    Quantifying autonomous vehicles national fuel consumption impacts: a data-rich approach

    Transp. Res. Part A Policy Pract.

    (2017)
  • D. Collingridge

    The Social Control of Technology

    (1980)
  • M.L. Cummings

    Integrating ethics in design through the value-sensitive design approach

    Sci. Eng. Ethics

    (2006)
  • G. De Melo et al.

    Multilingual text classification using ontologies

    29Th European Conference on Information Retrieval

    (2007)
  • T. Denning et al.

    Computer security and the modern home

    Commun. ACM

    (2013)
  • T. Denning et al.

    CPS: beyond usability: applying value sensitive design based methods to investigate domain characteristics for security for implantable cardiac devices

    30th Annual Computer Security Applications Conference

    (2014)
  • S. De Spiegeleire et al.

    Towards foresight 3.0: the HCSS metafore approach—A multilingual approach for exploring global foresights

  • I. Doxas et al.

    The dimensionality of discourse

    Proc. Natl. Acad. Sci. U. S. A.

    (2010)
  • N. Evangelopoulos et al.

    Latent semantic analysis: five methodological recommendations

    Eur. J. Inf. Syst.

    (2012)
  • A.K. Firat et al.

    Technological forecasting–A review. Working Paper Composite Information Systems Laboratory (CISL) 2008-15

    (2008)
  • M. Flanagan et al.

    Values at Play in Digital Games

    (2014)
  • J. Fleetwood

    Public health, ethics, and autonomous vehicles

    Am. J. Public Health

    (2017)
  • B. Friedman et al.

    Value Sensitive Design: Theory and Methods (UW CSE Technical Report 02-12-01)

    (2002)
  • R. Gheorghiu et al.

    Web 2.0 and the emergence of future oriented communities

    Econ. Comput. Econ. Cybern. Stud. Res.

    (2009)
  • P. Gray et al.

    From hindsight to foresight: applying futures research techniques in information systems

    Commun. Assoc. Inf. Syst.

    (2008)
  • Cited by (15)

    • Measuring topic network centrality for identifying technology and technological development in online communities

      2021, Technological Forecasting and Social Change
      Citation Excerpt :

      However, this approach ignores the phenomenon that new and emergent technology may be breakthrough technology, which only slightly depends on previous patent outputs, thus leading to objectivity "failure" when used to monitor emerging technologies. A third type of methodology is patent/literature-based text mining, such as text clustering (Lee et al., 2020; Zhou et al., 2019; Martin and Hüseyin, 2019; Kim et al., 2020; Kwon and Park, 2018), co-occurrence analysis (Diego et al., 2019; Dotsika and Watkins, 2017; Jaewoo, Woonsun, 2014; Ravikumar et al., 2015), topic modelling (Chen et al., 2017; Erzurumlu and Pachamanova, 2020; Jeong et al., 2019; Chen et al., 2020; Qiu and Wang, 2020), machine learning (Lee et al., 2018; Kristjanpoller and Minutolo, 2018; Kim and Sohn, 2020; Xu et al., 2019; Noh and Lee, 2020) and bibliometric analysis (Merino, 1990; Zhao et al., 2020), which are performed to identify technological topics and forecast development trends. The text-mining method to identify emerging technologies provides insights beyond patent data evaluation, reveals the transparency of patent structures, and may enhance the accuracy of identifying emerging technologies.

    • A review of data analytics in technological forecasting

      2021, Technological Forecasting and Social Change
      Citation Excerpt :

      Such websites are characterised by large participation and collaboration from many experts and the public (Kim et al., 2016a). Thus, they are regarded as valuable sources from which to analyse collective knowledge about technology trends and impacts; social, market, political changes; and future scenarios (Kwon and Park, 2018; Kwon et al., 2017). Fifthly, Wikipedia databases have been employed to identify the components of a technology (Kwon et al., 2018) and to measure the relationships between technologies at the macro-level (Cunningham, 2009).

    View all citing articles on Scopus
    View full text