Arkaitz Zubiaga

Arkaitz Zubiaga is a Senior Lecturer (Associate Professor) at Queen Mary University of London, where he co-leads the Social Data Science lab and serves as Director of Graduate Studies. He is also part of the leadership team of the Centre for Human-Centred Computing. His research sits at the intersection of Computational Social Science and Natural Language Processing, focusing on developing NLP and LLM methods for processing social media and Web data to tackle societal harms.

Zubiaga's research interests concentrate on addressing problematic issues with damaging societal effects, including hate speech, misinformation, inequality, biases, and other forms of online harm. He investigates how LLMs can be misused for malicious purposes such as spreading misinformation, generating abusive content, or exacerbating societal biases. His work emphasizes detecting and addressing irresponsible AI use where content is falsely claimed to be human-generated.

His publication record shows a clear trend toward addressing bias in detection systems, particularly in cyberbullying detection where swearing bias has been identified as a critical issue. His recent work explores zero-shot and few-shot learning approaches for cross-lingual applications, stance detection, and claim verification. The research spans multiple disciplines including computational linguistics, social computing, and AI ethics, with a growing focus on multimodal approaches and longitudinal model evaluation.

• 2024 OSNEM best survey award for work on session-based cyberbullying detection

Zubiaga actively mentors PhD students, with Peiling Yi recently passing her viva (April 2025) and welcoming new PhD students Alaa Bazaid and Ali Khairallah. He serves as senior area chair for ACL 2025 and leads the HYBRIDS MSCA Doctoral Network. His work demonstrates a strong commitment to developing responsible AI systems that can detect and mitigate online harms while addressing critical issues of bias and fairness in computational approaches.