Negative Feedback Fuels Hate Speech: A Deep Learning Analysis of 25 Million News Comments (Journalism Research News)

Hyo-sun Ryu and Jae Kook Lee of Sungkyunkwan University used deep learning hate speech detection software to examine 25 million comments made during the 2022 South Korean presidential election in their study ‘”‘Negative Feedback Fuels Hate Speech: A Deep Learning Analysis of 25 Million News Comments.'”‘ Seven theories were investigated: (1) people remark more when they receive positive feedback; (2) people who are not hateful post more when they receive positive feedback; (3) they post less when they receive negative feedback; and (4) and (5) hateful commenting increases when they receive both positive and negative criticism. Whether feedback effects differ for mild, medium, and high hate commentators was the focus of the sixth and seventh hypotheses. The work demonstrated efficient deep learning classification using examples of hateful and non-hateful comments. The results show that users react to input in general; non-hateful comments increased in response to positive feedback and decreased in response to negative criticism. Negative feedback, particularly from light hate commenters, boosted hateful posting whereas positive feedback had little effect. Contrary to certain assumptions, only comments that were not nasty demonstrated a definite beneficial correlation with response.

https://journalismresearchnews.org/article-negative-feedback-fuels-hate-speech-a-deep-learning-analysis-of-25-million-news-comments/