View Proposal

---

Proposer

Gavin Abercrombie

Title

Counterspeech Generation for Online Toxicity

Goal

Evaluate and develop datasets, methods, or systems to generate counterspeech responses to toxic comments

Description

Toxic and abusive language is increasingly prevalent in online spaces, but blocking and removal of such content is not always the best solution. An alternative is to automatically generate counterspeech to mitigate the effects of the abusive behaviour [1]. Recently, several datasets and approaches have been created for this [1, 2]. However, there remain many open questions about how best to generate and evaluate such responses [1, 2]. This project will investigate countermeasures to effectively fight the increasing online abuse without blocking freedom of speech. In particular, it will examine Natural Language Generation (NLG) algorithms to automatically generate counterspeech – responses that provide non-negative feedback through fact-bound arguments and broader perspectives to mitigate hate speech while fostering more harmonious conversations on social platforms. The project will focus on one or more of the following aspects of counterspeech: Analysis and evaluation of datasets; NLG for counterspeech responses; Evaluation of generated responses; Multilingual counterspeech

Resources

[1] Helena Bonaldi, Yi-Ling Chung, Gavin Abercrombie, and Marco Guerini. 2024. NLP for Counterspeech against Hate: A Survey and How-To Guide. In Findings of the Association for Computational Linguistics: NAACL 2024, pages 3480–3499, Mexico City, Mexico. Association for Computational Linguistics. [2] Chung, Y-L., Abercrombie, G., Enock, F., Bright, J. and Rieser, V. 2023. Understanding Counterspeech for Online Harm Mitigation.

Background

Url

Difficulty Level

Moderate

Ethical Approval

Full

Number Of Students

2

Supervisor

Gavin Abercrombie

Keywords

nlp, natural language generation, data analysis, counterspeech

Degrees

Master of Science in Artificial Intelligence
Master of Science in Artificial Intelligence with SMI
Master of Science in Data Science