Speaker
Description
Cyberbullying remains a critical online issue, often causing significant harm to victims. Although bystander intervention can help reduce its impact, promoting direct intervention remains challenging due to factors such as the effort required and limited confidence in expressing support. To address these barriers, this study developed EmojiGen, an innovative tool powered by a Large Language Model (LLM) designed to support bystanders in direct intervention efforts. To evaluate its effectiveness, we created a simulated social media platform featuring nine posts containing cyberbullying comments. A mixed-methods, between-subjects experiment with 90 participants participants was conducted to assess EmojiGen's impact on intervention behaviors and bystander perceptions. The results indicated that EmojiGen significantly increased supportive and resisting behaviors among bystanders. Additionally, it enhanced self-efficacy and intervention skills and reduced perceived workload and anxiety. This study provides valuable insights for fostering direct bystander intervention in cyberbullying and informs the design of supportive intervention tools for social networks.
Keywords
Cyberbullying Bystanders; Intervention; LLM-powered Assistance Tool
| Please indicate what type of scientific contribution it is | Mixed method study |
|---|---|
| Please also indicate what kind of contribution it is: | Scientific |
