TY - GEN
T1 - Towards Automated Detection of Risky Images Shared by Youth on Social Media
AU - Park, Jinkyung
AU - Gracie, Joshua
AU - Alsoubai, Ashwaq
AU - Stringhini, Gianluca
AU - Singh, Vivek
AU - Wisniewski, Pamela
N1 - Funding Information:
This research is supported in part by the U.S. National Science Foundation under grants #IIP-1827700, #IIS-1844881, and by the William T. Grant Foundation grant #187941. Any opinions, fndings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily refect the views of the research sponsors.
Publisher Copyright:
© 2023 ACM.
PY - 2023/4/30
Y1 - 2023/4/30
N2 - With the growing ubiquity of the Internet and access to media-based social media platforms, the risks associated with media content sharing on social media and the need for safety measures against such risks have grown paramount. At the same time, risk is highly contextualized, especially when it comes to media content youth share privately on social media. In this work, we conducted qualitative content analyses on risky media content flagged by youth participants and research assistants of similar ages to explore contextual dimensions of youth online risks. The contextual risk dimensions were then used to inform semi- and self-supervised state-of-the-art vision transformers to automate the process of identifying risky images shared by youth. We found that vision transformers are capable of learning complex image features for use in automated risk detection and classification. The results of our study serve as a foundation for designing contextualized and youth-centered machine-learning methods for automated online risk detection.
AB - With the growing ubiquity of the Internet and access to media-based social media platforms, the risks associated with media content sharing on social media and the need for safety measures against such risks have grown paramount. At the same time, risk is highly contextualized, especially when it comes to media content youth share privately on social media. In this work, we conducted qualitative content analyses on risky media content flagged by youth participants and research assistants of similar ages to explore contextual dimensions of youth online risks. The contextual risk dimensions were then used to inform semi- and self-supervised state-of-the-art vision transformers to automate the process of identifying risky images shared by youth. We found that vision transformers are capable of learning complex image features for use in automated risk detection and classification. The results of our study serve as a foundation for designing contextualized and youth-centered machine-learning methods for automated online risk detection.
KW - Instagram
KW - Private Message
KW - Self-supervised Learning
KW - Semi-supervised Learning
KW - Vision Transformer
KW - Youth Online Risk
UR - http://www.scopus.com/inward/record.url?scp=85159587915&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85159587915&partnerID=8YFLogxK
U2 - 10.1145/3543873.3587607
DO - 10.1145/3543873.3587607
M3 - Conference contribution
AN - SCOPUS:85159587915
T3 - ACM Web Conference 2023 - Companion of the World Wide Web Conference, WWW 2023
SP - 1348
EP - 1357
BT - ACM Web Conference 2023 - Companion of the World Wide Web Conference, WWW 2023
PB - Association for Computing Machinery, Inc
T2 - 2023 World Wide Web Conference, WWW 2023
Y2 - 30 April 2023 through 4 May 2023
ER -