Understanding Generative AI Risks for Youth: A Taxonomy Based on Empirical Data
Project Overview
The document explores the significant role of Generative AI (GAI) in education, detailing its transformative effects on youth's engagement with technology while identifying a comprehensive taxonomy of 84 associated risks, categorized into six main themes, including mental wellbeing and behavioral development. These risks arise from the dynamic and adaptive nature of interactions with GAI, potentially resulting in serious issues such as cyberbullying, addiction, and privacy violations. The authors stress the importance of awareness among stakeholders—such as educators and policymakers—to recognize and address these risks to ensure the responsible integration of GAI in educational settings. Ultimately, the findings underscore the dual-edged nature of GAI's potential to enhance learning experiences while simultaneously necessitating careful consideration of the risks involved.
Key Applications
Youth-GAI Interaction Risk Taxonomy
Context: Youth interactions with Generative AI tools such as chatbots, particularly in educational settings.
Implementation: Developed a taxonomy based on empirical data from chat logs, Reddit discussions, and AI incident reports.
Outcomes: Provides a structured understanding of the risks youth face, helps in designing safer AI systems, and informs policymakers on regulations.
Challenges: The taxonomy may not capture all youth-GAI interactions across diverse cultural, linguistic, and socioeconomic backgrounds.
Implementation Barriers
Awareness Barrier
Parents and educators may lack awareness of the specific risks associated with youth interactions with GAI.
Proposed Solutions: Develop educational resources and training for parents and educators to understand the risks and safe practices for youth-GAI interactions.
Design Barrier
Existing GAI systems may not have adequate safety features or moderation tools to protect youth. It is essential to implement more robust moderation mechanisms that adapt to the evolving nature of youth-GAI interactions.
Proposed Solutions: Implement more robust moderation mechanisms that adapt to the evolving nature of youth-GAI interactions.
Project Team
Yaman Yu
Researcher
Yiren Liu
Researcher
Jacky Zhang
Researcher
Yun Huang
Researcher
Yang Wang
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Yaman Yu, Yiren Liu, Jacky Zhang, Yun Huang, Yang Wang
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