I am Marco Postiglione, a Postdoctoral Research Scholar in the Northwestern Security & AI Lab at Northwestern University, where I work with Prof. V. S. Subrahmanian. My research spans AI security, biomedical informatics, and trustworthy machine learning, with a commitment to developing AI systems that advance social good. Much of my current work focuses on building reliable, scalable, and adversary-resilient AI technologies for global media integrity, public safety, and health applications.

At Northwestern, I help lead several large applied-AI initiatives. I deployed the Global Online Deepfake Detection System (GODDS), now used by 70+ news organizations and fact-checkers for real-time AI-generated media verification. I also co-developed the MNW Benchmark for deepfake detection in partnership with Microsoft AI for Good Lab and WITNESS, and designed a context-aware audio deepfake detector that substantially improves over state-of-the-art systems under adversarial conditions.

Beyond media forensics, I work on SMART (Social Movement Analysis & Reasoning Tool), built in collaboration with journalists from The Wall Street Journal, The Washington Post, The Associated Press, and others to analyze discourse dynamics and event relationships in social movements such as #MeToo and Black Lives Matter. I also contribute to national security applications, including DEWS (a drone early-warning system deployed with the Netherlands Police and Municipality of The Hague) and the Northwestern Terror Early Warning System (NTEWS).

I earned my Ph.D. in Computer Science and Engineering from the University of Naples Federico II, advised by Prof. Vincenzo Moscato. My doctoral research focused on AI for healthcare, including few-shot learning for Italian clinical text, biomedical NLP, and temporal knowledge graph models for predicting disease progression. As a Visiting PhD Researcher at King’s College London in the Department of Biostatistics & Health Informatics, I developed a temporal knowledge graph framework integrating clinical histories with medical ontologies to forecast future disorders.

📧 Feel free to reach out to me at my email address: marco (dot) postiglione (at) northwestern (dot) edu

Recent news

🟢 [Nov 2025] Our paper “A Nonpartisan Study of Deepfake Activity and Engagement Around the 2024 US Presidential Election” has been accepted to the International AAAI Conference on Web and Social Media (ICWSM 2026)!

🟢 [Nov 2025] “DEEP: A Discourse Evolution Engine for Predictions about Social Movements” has been accepted to AAAI 2026 (IAAI Track). Thrilled to continue supporting work at the intersection of computational social science and journalism.