Algorithmic decision-making in social networks

PI: Ms. Lital Kahalon Zandberg. 

Abstract:
Can algorithmic enforcement effectively manage violent speech in social networks?

Social media offer a platform for users to communicate, exchange and distribute content. Users are the ones who determine the nature of use of the social media, deciding which information to make available to the public and how. Some content posted on social media platforms might be inciting, threatening or otherwise harmful. It is quite challenging to monitor potentially harmful contents and prevent its publication when necessary. The network is not limited in time or space, and harmful content could become promptly accessible and widely distributed, thereby amplifying its potential harm. The core features of the network are not only viral but also eternal, and consequently, any harmful content that was uploaded to the internet will remain there, unless an active act to removal is taken.

Over the past decade, there has been an increasing use of algorithms for enforcement purposes. Algorithms are used as a tool for monitoring, filtering, blocking and removing content.

The purpose of this research is to examine the efficiency of using algorithms on social media for the purpose of enforcement and policing. The research will also examine how algorithms adapt to changes by machine learning, thus improving its result in accordance to the details of users.

If there are any offenses made or damages caused on social network platforms, liability may be imposed by using the guidance of the existing laws. So far, there is no clear statutory framework governing algorithmic enforcement, therefore a research examining the possibility of regulating behavior in social networks by means of algorithmic enforcement is required. This research will focus on applying algorithmic enforcement in social networks in the context of cyberbullying.