Title: Tackling Online Threats to Democracy in High-Stakes Low-Resource Contexts

Date: November 28th, 2023

Time: 10:00 AM - 12:00 PM EST

Location: Hybrid

In-person: CODA C1015 Vinings

Virtual: Join Zoom meeting 

https://gatech.zoom.us/j/98346182572?pwd=UG9wemsyZWRGTGVLd2V2VmNyaUNxUT09

Daniel K. Nkemelu

Ph.D. student in Human-Centered Computing

School of Interactive Computing

Georgia Institute of Technology

Committee:

Dr. Michael L. Best (co-advisor) – School of Interactive Computing and Sam Nunn School of International Affairs, Georgia Institute of Technology

Dr. Irfan Essa (co-advisor) – School of Interactive Computing, Georgia Institute of Technology

Dr. Ellen Zegura – School of Computer Science, Georgia Institute of Technology

Dr. Munmun De Choudhury – School of Interactive Computing, Georgia Institute of Technology

Dr. Neha Kumar – School of Interactive Computing and Sam Nunn School of International Affairs, Georgia Institute of Technology

Dr. Monojit Choudhury - Turing India, Microsoft

Abstract

The rapid adoption of social media around the world has been accompanied by a surge in the dissemination of problematic content such as hate speech, disinformation, and misinformation. Social media platforms and regulators have historically overlooked the potentially devastating effects of these contents in non-Western contexts. Consequently, the burden of preserving online integrity falls on civil society actors who are often overworked, under-resourced, and underappreciated. With the relative success of machine learning applications in research and practice, there is an urgent need to understand how these technologies can support civil society organizations working to tackle these problems and build safer, more resilient democratic societies.

My dissertation focuses on two types of civil society actors: social media monitors and fact-checkers. Using a mixed-method approach, my research investigates how machine learning can effectively support these stakeholders' efforts to track and respond to hate speech, disinformation, and misinformation in low-resource contexts. In my completed work, I have partnered with social media monitors in Myanmar to source and label data, and train and deploy machine learning models to support hate speech monitoring during the 2020 Myanmar elections. I have drawn on lessons from this work to develop a contextual entity substitution method for hate speech data augmentation in limited data contexts. 

In my proposed work, I seek to center the perspectives of fact-checkers working to address online misinformation and disinformation. My research aims to understand their current strategies and tools for creating and disseminating fact-checked content, how users respond to published fact-checked content, and how the development of an AI-driven tool for making fact-checks accessible can reduce the burden and amplify the impact of their work. This dissertation will advance our understanding of online threats to democracy in the Global South and provide insights that improve the capacities of stakeholders to respond to these challenges.

Thank you,

Daniel Nkemelu

Daniel Nkemelu | dnkemelu@gatech.edu

PhD Student,

School of Interactive Computing,

Georgia Institute of Technology

From: Nkemelu, Daniel K <dnkemelu@gatech.edu>
Sent: Friday, November 24, 2023 3:18 PM
To: phd-coc-announce@cc.gatech.edu <phd-coc-announce@cc.gatech.edu>; faculty@cc.gatech.edu <faculty@cc.gatech.edu>; announcements@grad.gatech.edu <announcements@grad.gatech.edu>
Cc: Best, Michael <mikeb@gatech.edu>; Essa, Irfan A <irfan@gatech.edu>; Zegura, Ellen <ewz@cc.gatech.edu>; De Choudhury, Munmun <munmun.choudhury@cc.gatech.edu>; Kumar, Neha <neha.kumar@gatech.edu>; Monojit Choudhury <monojitc@microsoft.com>; Nash, Theresa L <tnash33@gatech.edu>
Subject: PhD Thesis Proposal Announcement

Dear all,

You are cordially invited to my thesis proposal on Tuesday, November 28th. 

Title: Tackling Online Threats to Democracy in High-Stakes Low-Resource Contexts

Date: November 28th, 2023

Time: 10:00 AM - 12:00 PM EST

Location: Hybrid

In-person: CODA C1015 Vinings

Virtual: Join Zoom meeting 

https://gatech.zoom.us/j/98346182572?pwd=UG9wemsyZWRGTGVLd2V2VmNyaUNxUT09

Daniel K. Nkemelu

Ph.D. student in Human-Centered Computing

School of Interactive Computing

Georgia Institute of Technology

Committee:

Dr. Michael L. Best (co-advisor) – School of Interactive Computing and Sam Nunn School of International Affairs, Georgia Institute of Technology

Dr. Irfan Essa (co-advisor) – School of Interactive Computing, Georgia Institute of Technology

Dr. Ellen Zegura – School of Computer Science, Georgia Institute of Technology

Dr. Munmun De Choudhury – School of Interactive Computing, Georgia Institute of Technology

Dr. Neha Kumar – School of Interactive Computing and Sam Nunn School of International Affairs, Georgia Institute of Technology

Dr. Monojit Choudhury - Turing India, Microsoft

Abstract

The rapid adoption of social media around the world has been accompanied by a surge in the dissemination of problematic content such as hate speech, disinformation, and misinformation. Social media platforms and regulators have historically overlooked the potentially devastating effects of these contents in non-Western contexts. Consequently, the burden of preserving online integrity falls on civil society actors who are often overworked, under-resourced, and underappreciated. With the relative success of machine learning applications in research and practice, there is an urgent need to understand how these technologies can support civil society organizations working to tackle these problems and build safer, more resilient democratic societies.

My dissertation focuses on two types of civil society actors: social media monitors and fact-checkers. Using a mixed-method approach, my research investigates how machine learning can effectively support these stakeholders' efforts to track and respond to hate speech, disinformation, and misinformation in low-resource contexts. In my completed work, I have partnered with social media monitors in Myanmar to source and label data, and train and deploy machine learning models to support hate speech monitoring during the 2020 Myanmar elections. I have drawn on lessons from this work to develop a contextual entity substitution method for hate speech data augmentation in limited data contexts. 

In my proposed work, I seek to center the perspectives of fact-checkers working to address online misinformation and disinformation. My research aims to understand their current strategies and tools for creating and disseminating fact-checked content, how users respond to published fact-checked content, and how the development of an AI-driven tool for making fact-checks accessible can reduce the burden and amplify the impact of their work. This dissertation will advance our understanding of online threats to democracy in the Global South and provide insights that improve the capacities of stakeholders to respond to these challenges.