Explanations as Mechanisms for Supporting Algorithmic Transparency

Abstract

Transparency can empower users to make informed choices about how they use an algorithmic decision-making system and judge its potential consequences. However, transparency is often conceptualized by the outcomes it is intended to bring about, not the specifics of mechanisms to achieve those outcomes. We conducted an online experiment focusing on how different ways of explaining Facebook's News Feed algorithm might affect participants' beliefs and judgments about the News Feed. We found that all explanations caused participants to become more aware of how the system works, and helped them to determine whether the system is biased and if they can control what they see. The explanations were less effective for helping participants evaluate the correctness of the system's output, and form opinions about how sensible and consistent its behavior is. We present implications for the design of transparency mechanisms in algorithmic decision-making systems based on these results.

Playing the visibility game: How digital influencers and algorithms negotiate influence on Instagram

“…Indeed, some users are not aware of algorithmic ranking on platforms (Eslami et al, 2015;Rader et al, 2018) and, thus, do not even know the game exists, let alone how to play it. This suggests that the game reinforces offline hierarchies of social privilege with 'winners' being those with greater access to social, cultural, political, and economic resources.…”

“…Like other platform owners, Instagram shares few details about the platform's algorithmic architecture or how it works. In general, algorithms function behind the scenes with many users unaware of their presence (Eslami et al, 2015;Rader et al, 2018). Even so, algorithms play an important role in structuring our online experiences (Beer, 2009;Bucher, 2012;Cheney-Lippold, 2011;Gillespie, 2014).…”

Agent, Gatekeeper, Drug Dealer

“…Prior research has found that explanations often led to more awareness of algorithmic decisions and helped users determine whether a system was biased. It also changed their behaviors toward the system and increased their confidence [16,68]. The possibility of an appeal process for algorithmic systems has received less attention in the literature but would be a rich area for future work.…”

Ethical Dimensions for Data Quality

“…Transparency is the ability to interpret the information extraction process in order to verify which aspects of the data determine its results. In this context, transparency metrics can use the notions of (i) data provenance [19,18], by measuring the amount of meta-data describing where the original data come from; (ii) explanation [15], by describing how a result has been obtained. Diversity is the degree to which different kinds of objects are represented in a dataset.…”