Översikt

  • Unit 3: Sociotechnical institutions: Bias in AI and algorithms

    The influence of AI has become a powerful institutional force on society because algorithms reinforce and perpetuate existing structures or introduce new patterns with far-reaching consequences. Machine learning algorithms and AI are increasingly used in organizations. They make decisions that have important consequences for individuals, such as hiring and loan applications. AI learns from statistical co-occurrences in past data sets, which means that they are prone to perpetuate historical biases. In this unit, we learn about how and why this contemporary issue occurs and learn about some suggestions of how it might be solved (Morse, Teodorescu, Awwad, & Kane, 2020).

    In the group discussion of this unit, your primary task is to discuss some or all of the guiding questions that are listed in the course syllabus. Further, you can also use the Scott (2014) three pillars framework as a theoretical framework for analysis and identify elements of the pillars in the other readings. AI and algorithms have been included in this course because they constrain, enable and orient human behavior and actions, they can be widespread, have real consequences, are frequently taken for granted, and are made by humans. In the group discussion or your essay, you can, if you want to, also discuss whether AI and algorithms can be seen as a fourth technological pillar of institutions and what a fourth pillar would look like in Scott’s (2014) framework. There is no right answer to this question.


    Article (academic, media or other)

    Why this article / what part of the article

    Morse, L., Teodorescu, M., Awwad, Y., & Kane, G. C. (2020). A Framework for Fairer Machine Learning in Organizations. Available at SSRN 3690570.

    Very recent article presented at the AOM2020 conference that proposes solutions as well. Read the full article. (32 pages, double-spaced)

    Z_ai (2020). Bias in artficial intelligence. Towardsdatascience.com.  https://towardsdatascience.com/bias-in-artificial-intelligence-a3239ce316c9

    Technology blog on bias in AI. Read the full blog post. (8 pages)

    Dastin (2018). Amazon scraps secret AI recruiting tool that showed bias against women. Reuters.  https://www.reuters.com/article/us-amazon-com-jobs-automation-insight-idUSKCN1MK08G

    Illustrative example of issues caused by biased algorithms. Voluntary reading.

    Scott, W. R. (2014). Institutions and organizations: Ideas, interests, and identities. Sage publications. Pages 59-70.

    The first edition was published in 1995.

    Voluntary article as a theoretical framework for analysis or conceptual development of a four pillar theoretical framework.

     



    • Mapp icon
      Unit 3: Readings Mapp
    • Forum icon
      Unit 3 Discussion forum
    • Inlämningsuppgift icon
      Unit 3 Group discussion report Inlämningsuppgift
    • Inlämningsuppgift icon
      Unit 3 List of group discussion attendants Inlämningsuppgift
    • Turnitin Uppgift icon
      Unit 3 Individual essay Turnitin Uppgift
    • Mapp icon
      Illustrative examples of statistical bias Mapp
◄ Previous section
Unit 2: Shareholders, stakeholders, and corporate control
Next section ►
Unit 4: Sustainability institutions: Reporting, certification, blockchain tracing, and GHG accounting scopes