Natural Language Processing (NLP) has become one of the most influential areas of computer science, impacting not only technology but also society as a whole. Yet beneath the rapid progress lies a set of structural dynamics. In this talk, I take the audience into some parts “under the hood” of the NLP research ecosystem by drawing on multiple large-scale empirical studies that interrogate the field’s dynamics. Specifically, I look into stakeholders, cross-field influence, citation practices, peer review, and biases. I highlight the growing presence of industry in NLP publications, demonstrating how a few large companies have substantial research outputs and claim a high number of citations. I examine how NLP engages with other disciplines. Although NLP research once bridged fields, cross-disciplinary citation diversity is at an all-time low, with the field increasingly citing itself. In addition, we see a growing recency bias that highlights how rapidly the field is evolving at the expense of works older than just a few years. Peer review faces substantial challenges due to a rapid increase in papers and pressure on organizers and reviewers. There are still clear disadvantages for marginalized groups and “glass ceilings” that hinder them from fully implementing their potential. These findings, along with those of related studies in the field, raise broader questions about the direction the field is taking and what we can do to gain agency. To facilitate future research, I also present large corpora for papers, books, social media data, and AI-generated content that can help answer new research questions about NLP and computational affective science, connecting NLP with psychology, sociology, and cognitive science.
Invited Speaker: Jan Philip Wahle (University of Göttingen)
Bio: Jan Philip Wahle is a research fellow in computer science at the University of Göttingen in Germany. He received his Master’s degree in computer science from the University of Wuppertal and worked for the automotive company Aptiv PLC before continuing with his Ph.D. studies. During his Ph.D., Jan has been a visiting researcher at the National Research Council of Canada. His research has been presented at ACL, EMNLP, EACL (among others). One of his research interests lies in gaining a deeper understanding of the dynamics of the NLP research field from a meta-perspective, particularly in identifying certain tendencies and biases within it. These days, he also manages third-party projects to advance low-resource models for historical German texts and projects related to AI safety and interpretability.