Using Automated Digital Systems to Thoroughly Regulate Social Governance: Monitoring and Behavior Modification through Data-driven Algorithmic Decision-Making.

• Author: Androniceanu, Armenia

1. Introduction

   Both prediction and decision-making are mainly calculative operations that can be adequately automated through algorithmic means. (Fazi, 2018) Algorithms conjecture user choices and recommend content, but customization may regulate individuals' socio-economic involvement and possibly their belief systems. (Olhede and Rodrigues, 2017) Assessments derived from algorithmic, machine learning patterns can be discriminatory, replicating partisanships in historical input employed to train them. (Veale and Binns, 2017)

2. Conceptual Framework and Literature Review

   In the current information society, tasks and assessments are progressively redistributed to machines, automated systems, and artificial agents that facilitate human connections, by reaching decisions and proceeding by means of algorithms. (D'Agostino and Durante, 2018) Computation cannot generate originality as algorithmic formalization is incapable of creating any knowledge, instead disintegrating it into a limited and discrete handling of symbols and guidelines. (Fazi, 2018) Machine learning employs historical statistics to guide the algorithm what aspects to include so that to attain a certain objective. (Martin, 2018) Algorithmic decision-making processes may result in more non-discriminatory and consequently possibly more impartial assessments than those performed by human beings (Lepri et al., 2018). Algorithms may disfavor in the wake of a social category, deliberately or not, even without being unequivocally fed social category information. (Williams et al., 2018)

3. Methodology and Empirical Analysis

   Building my argument by drawing on data collected from HubSpot, Pew Research Center, and Statista, I performed analyses and made estimates regarding % of users who say they frequently/sometimes see content on social media that makes them feel amused/angry/connected/inspired/depressed/lonely, % of users who say they more often see people being mean or bullying/kind or supportive/trying to be deceptive/trying to point out inaccurate info when using social media sites, and % of users who say it would be very/somewhat difficult/easy for social media sites to figure out their race or ethnicity/hobbies and interests/political affiliation/religious beliefs. Data collected from 4,800 respondents are tested against the research model by using structural equation modeling.

4. Results and Discussion

   Algorithms are designed to reorganize, reconstruct and mature like living organisms. (Fazi, 2018) Algorithmic mechanisms and applications should be inspected and governed carefully, if individuals aim to hinder the temptations of information and communications...