Data-Driven
Cybersafety Lab
We develop methods to study, measure, and mitigate harms on online platforms. Our work lies at the intersection of computer security, artificial intelligence, and computational social science.
4+
Years Active
5
Researchers
30+
Publications
Latest News
View all news →- Happy to share that our paper entitled “Expectation, Backlash, Recovery, and Excitement: How Model Releases Shape Reddit Perceptions of Conversational AI Systems,” has been accepted at EMNLP 2026 (Main Conference)! 🎉 Huge congratulations to Vahid Rahimzadeh for leading the effort and for getting his first PhD paper accepted at a top NLP venue! Many thanks also to Yury Zhauniarovich for being an amazing co-supervisor and collaborator!
- Happy to share that our paper entitled “AudiTok: A System for Cross-Country Audits of TikTok’s Mobile App” has been accepted at ICWSM 2027! 🎉 Congratulations to Moonis for leading the effort!
- 🏆 We are thrilled to share that our PI, Savvas Zannettou, has been awarded the prestigious Adamic-Glance Distinguished Early Career Award 2026 at AAAI ICWSM 2026! This annual award recognizes a researcher who has distinguished themselves through creativity and rigor in addressing computational social science research topics of major societal impact — a tremendous honor for the lab and our collaborators.
- 🏆 Our paper 'The Great Data Standoff: Researchers vs. Platforms Under the Digital Services Act' received a Best Paper Honorable Mention Award at AAAI ICWSM 2026. Congratulations to Catalina and Thales for leading the efforts!
- 🏆 Our paper 'Gold Standard or Gold-Plated? Human Practices of Triple Verification in CSAM Takedown' received an Honorable Mention Award at ACM CHI 2026. Congratulations to the first author, Melissa!
Research Areas
Learn more →Online Harms & Platform Accountability
Measuring harmful content, recommender system effects, and platform interventions at scale. We build methods to hold platforms accountable to the public.
AI-Driven Cybersecurity & Cybersafety
Applying artificial intelligence to detect, model, and mitigate emerging threats in online environments.
Hate Speech & Content Moderation
Auditing moderation systems and developing fairer, more transparent approaches to detecting and removing harmful speech.