Dr Ana Jovančević

Dr Dr Ana Jovančević
Marie Sklodowska-Curie Fellow
University of Limerick
Profile
Ana Jovancevic, PhD, is a social experimental psychologist whose research bridges psychology, technology, and society. She earned her doctorate at the University of Niš, Serbia, where her thesis examined implicit biases toward individuals with obesity through innovative methods such as eye-tracking and information manipulation. Her findings demonstrated how unconscious prejudices emerge through subtle gaze patterns, offering fresh insights into the mechanisms of stereotype formation. Building on this foundation, Ana developed an international research profile through collaborations across Europe and wider with experts in psychology, computer science, and mathematics. After completing her PhD, she joined the University of Limerick as a postdoctoral researcher on the ERC-funded DAFINET project and later on the ERC-funded ID-COMPRESSION project. Her research there combined social psychology with network analysis, employing advanced quantitative methods to uncover how identity and attitudes shape complex social systems. To date, she has authored 18 peer-reviewed articles, and her work has received 298 citations on Google Scholar.
Beyond research, Ana is a dedicated teacher, mentor, and science communicator. She has taught a wide range of courses—including perception, psychometrics, empirical research in art, and social psychology—while supervising numerous BSc and MSc projects. Deeply committed to open science and societal impact, she served as editorial assistant for Psihologija journal and founded the Association for Integration of Research and Practice in Psychology to bridge academic knowledge with applied practice. She also served on multiple conference committees. Her achievements have been recognized through multiple awards for academic excellence. Through her interdisciplinary work, Ana continues to advance innovative approaches to understanding bias, identity, and human–AI interaction.
Contact Information
LinkedIn: https://www.linkedin.com/in/ana-jovan%C4%8Devi%C4%87-5b8888210/?originalSubdomain=ie
Google Scholar: https://scholar.google.com/citations?user=JydKAw0AAAAJ&hl=sr
ORCID ID: 0000-0002-1693-8891
ROSETTA Project
AI PRIVElege-by-AGE (AI-PRIVELAGE): Exploring age-related identity privileges in AI interactions
Artificial Intelligence (AI) based on Large Language Models (LLMs; e.g., ChatGPT, LLaMA) offers major benefits, but also risks favouring some groups while disadvantaging others, especially underrepresented ones like children and the elderly. While algorithmic bias (built into the models) is already well known, AI-PRIVELAGE identifies two new forms of bias that appear at the intersection of users’ social identities and model properties, in line with the EU’s goal of fair AI use:
- Identity pinning happens when a user’s social identity influences AI output quality through unintentional prompt engineering. For example, age-related language styles may change how queries are understood, producing less useful or lower-quality results for marginalised groups.
- Vulnerability bias assumes that children and the elderly are more exposed to AI-related risks, meaning they ‘pay’ an ‘AI identity tax’. With lower digital skills, they may be more vulnerable to misinformation (‘hallucinations’) and to impostor feelings, especially when negatively stereotyped groups face more suspicion of AI misuse.
AI-PRIVELAGE will combine fellow’s expertise in human-computer interaction, eye-tracking, implicit measures, and psychometrics with the host’s strengths in computational approaches to social identity theory, identity-based language analysis, and computer science. This collaboration will advance knowledge of life-stage-related AI biases and social inequalities.
ROSETTA Supervisor and Host Institution
Supervisor: Proff Mike Quayle
Host: University of Limerick