|
Projects
Detecting Adolescent Online Risks
We develop privacy-preserving, context-aware algorithms to improve the accuracy and fairness of online risk detection.
Jump Down:
Algorithmic approaches for detecting online risks
This research partnership, titled “A Multi-Disciplinary Approach to Detecting Adolescent Online Risks,” brings together social science, computer science, and industry collaboration to create algorithmic approaches for detecting online risks adolescents face. The project centers teen voices by grounding machine learning in real experiences, ensuring that tools developed to detect and flag risks are developmentally appropriate and contextually relevant.
Focusing on addressing cyberbullying, grooming, and self-harm, the team is developing privacy-preserving, context-aware algorithms to improve the accuracy and fairness of online risk detection. These efforts are informed by participatory research and aim to create both open-source tools and commercial applications for embedding safety directly into platforms.
By working closely with academic and industry partners, the project advances proactive, ethical AI systems that prioritize youth safety and trust.
Project Team
ICSI Research Team
Prakriti Dumaru
Postdoctoral Researcher
About
Focus Areas
- Human-AI Collaboration and Teaming
- Digital Wellbeing, Behavior-Aware Computing, and Ethical Design
- Social Media Platforms, Online Communities, and Digital Culture
Get in touch
Want to discuss opportunities to work with ICSI? We’d love to hear from you.