Generating A Crowdsourced Conversation Dataset to Combat Cybergrooming
Project Overview
The document focuses on the application of generative AI in education, particularly in addressing the issue of cybergrooming among adolescents. It describes the creation of a conversational agent designed to educate youth about the risks associated with cybergrooming, which poses significant threats to their mental health. By leveraging generative AI, the project aims to generate an authentic dataset that reflects real-life interactions, thereby improving the effectiveness of the educational tool. Crowdsourcing responses from both parents and adolescents is a key strategy employed to enrich the dataset, fostering a deeper understanding of the dangers of online predatory behaviors and enhancing coping mechanisms. This innovative approach not only raises awareness about the risks but also empowers young individuals and their guardians to recognize and respond to such threats effectively. The findings suggest that by utilizing generative AI in this capacity, educational tools can be more engaging and relevant, ultimately leading to better outcomes in safeguarding youth in the digital landscape.
Key Applications
Conversational agents for simulating conversations between adolescents and predators
Context: Educational context aimed at adolescents and parents
Implementation: Crowdsourcing conversation responses from parents and adolescents through an online survey in simulated cybergrooming scenarios
Outcomes: Empowers adolescents with knowledge about predatory behaviors and enhances coping mechanisms; provides insights into the differences in risk perception between parents and adolescents.
Challenges: Challenges include the lack of authentic datasets, privacy concerns, and the difficulty of accurately simulating real-world conversations.
Implementation Barriers
Data Access Barrier
Difficulty in accessing authentic datasets that reflect modern communication among adolescents.
Proposed Solutions: ['Propose to generate a new dataset by involving both parents and adolescents to capture a wider range of responses and behaviors.', 'Use diverse and updated datasets that capture authentic adolescent interactions and responses.']
Privacy Concerns
Monitoring online communications raises privacy issues for individuals.
Proposed Solutions: Implement ethical considerations and recruitment strategies that protect participants' identities and privacy.
Behavioral Differences
Adolescents in real-world situations may behave differently than simulated participants.
Proposed Solutions: Use diverse and updated datasets that capture authentic adolescent interactions and responses.
Project Team
Xinyi Zhang
Researcher
Pamela J. Wisniewski
Researcher
Jin-hee Cho
Researcher
Lifu Huang
Researcher
Sang Won Lee
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Xinyi Zhang, Pamela J. Wisniewski, Jin-hee Cho, Lifu Huang, Sang Won Lee
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai