Kristina Gligorić joins the Johns Hopkins University as an assistant professor of computer science and a member of the Data Science and AI Institute.
Gligorić received her PhD in computer science from the École Polytechnique Fédérale de Lausanne in 2022. Before joining Hopkins, she was a postdoctoral scholar in the Computer Science Department at Stanford University.
Tell us a little bit about your research.
My work sits at the intersection of AI, computational methods, and social sciences. I study how to build better natural language processing systems that understand human behavior, and the other way around—how to use them to better understand society. We ask questions like, How can we develop socially aware AI methods, drawing inspiration from the social sciences? Using AI for social applications, we’re asking questions such as, How can we apply these tools to help address urgent societal problems? and How can we develop tools that support scientific discovery in the social sciences, enabling us to better understand humans, our behaviors, and the decisions we make? Today’s off-the-shelf models are not designed to solve these kinds of challenges, so our research fills this gap by, for example, collecting specialized datasets and training models that can actually support practitioners’ real needs.
Tell us about a project you are excited about.
I’m particularly excited about a line of work focused on building AI tools that make large-scale data annotation feasible and scientifically reliable for social science research, with the broader goal of understanding and explaining human behavior. An early result in this direction is our recent paper, “Can Unconfident LLM Annotations Be Used for Confident Conclusions?”, which addresses a core challenge in computational social science: leveraging fast, inexpensive annotations from large language models without undermining the validity of downstream scientific conclusions. Published at the 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics, the paper introduces a framework that uses model uncertainty to determine when human annotations are needed, enabling researchers to combine LLM outputs with a small number of human labels while still producing statistically valid estimates and confidence intervals.
Building on these initial results, we are now developing approaches that support much harder annotation settings that are central to social science: multimodal data such as video and audio, inherently ambiguous tasks where reasonable annotators may disagree because they bring different perspectives, and adversarial settings where researchers may intentionally “hack” annotation pipelines to shape conclusions. Ultimately, our goal is to design AI-assisted annotation systems that not only scale data collection but also do so in ways that respect uncertainty, disagreement, and strategic behavior, making them suitable for rigorous studies of human behavior.
Why this? What drives your passion for your field?
My goal is to build technology that genuinely improves people’s lives, helping us learn new skills, communicate more effectively, and make better decisions. But at the same time, new technologies—as we’re seeing with LLMs—also create a lot of risks. I work at this exact intersection, understanding how people behave and how technology shapes those behaviors while designing AI systems that amplify benefits and minimize harm.
What excites me the most is that LLMs that can better understand our communication and generate fluent text offer new opportunities to make progress on foundational social problems. We can understand large amounts of data much faster, we can generate hypotheses about what drives our behaviors and decisions, we can help people who are trying to positively impact our behavior, and we can even predict how we will react to interventions or changes in the environment. This is a very exciting time to be working in this area.
What classes are you teaching?
This semester, I’m teaching Ethics of AI and Automation. The goal of this class is to familiarize students with foundational concepts around the ethical development of AI, key risks, and best development practices to minimize these risks. It’s a discussion-based class, so we will read and discuss research papers, media reports, and policy documents. Students will also work on a project focusing on specific ethical risks of AI systems. In the fall, I will teach an advanced graduate class focusing on advanced topics in social natural language processing. Both of these courses connect to my research, focusing on building language technologies that understand human behavior and ensuring that new technologies are developed and deployed responsibly.
Why are you excited to be joining the Johns Hopkins Department of Computer Science?
It’s incredibly exciting to be joining the department, which has long been known as a leader in the development and applications of language technologies. From day one I’ve felt at home interacting with faculty working in this area. In addition to this expertise and reputation around language, the historical growth in data science and AI here is incredible. It’s great to be part of that momentum; it’s a truly unique opportunity to shape how we develop AI and steer its development toward positive impact, building on existing expertise across all areas of computer science. Johns Hopkins provides an ideal environment for the interdisciplinary work I do, with proximity to leading global experts in any field you can think of across medicine, public health, and the social sciences.
I’m also very excited to be joining a university that is invested in making education broadly accessible, especially through free tuition initiatives for the majority of families. On my journey, I’ve personally benefited from accessible public education at all levels of education, and it feels great to support this community through my own research, teaching, and mentoring.
Besides your work, what are some of your other hobbies and passions?
I love traveling, seeing new places, nature, learning about cultures, visiting museums, and learning about foods. I enjoyed a recent road trip in New Mexico where I got to try some amazing green chiles and sopapillas. I also love staying at home and recharging my battery! I love reading, cooking, and yoga. I’ve recently started playing ukulele, and have fun adapting catchy pop songs to this lovely little instrument.