Human-Centered AI · Human–AI Interaction
Luca
Marconi
I study how humans and AI can collaborate effectively while preserving judgment, expertise, agency, critical thinking, and responsibility.
Department of Informatics, Systems and Communication (DISCo) · MUDI Lab · University of Milano-Bicocca, Milan
Research at a glance
Three connected pillars.
One overarching question.
How can AI systems be evaluated, designed, and governed so that Human–AI collaboration strengthens rather than substitutes human capabilities?
Interaction Quality & Human–AI Evaluation
Evaluating cognitive, behavioral, collaborative, and experiential dimensions of Human–AI interaction beyond system performance alone.
Meaningful Human Oversight
Studying calibrated reliance, cognitive offloading, automation bias, deskilling, and the preservation of agency, competence, expertise, and responsibility.
Responsible AI Governance & Deployment
Connecting Human-Centered AI with validation, transparency, standardization, accountability, and trustworthy real-world deployment.
Core research assets
Frameworks and empirical projects
Representative frameworks and empirical projects illustrate how the broader research agenda is translated into methods, instruments, data, and real-world applications.
HADQ
Human–AI Dialogue Quality: a multidimensional framework and research toolkit for conversational AI evaluation.
Methodological frameworkCLARITY AI
Multidimensional evaluation of methodological quality, transparency, robustness, explainability, ethics, usability, and translational readiness in Medical AI.
Empirical projectHuman–AI Collaboration in Education
Structured roles and interaction protocols for LLM-supported learning, cognitive engagement, calibrated reliance, and responsible adoption.
Emerging directions
Emerging directions across the program
Complementary directions extending the current research program toward sustained human capabilities, adaptive co-agency, longitudinal effects, and richer Human–AI evaluation.
Cognitive Sustainability
How sustained interaction with AI shapes judgment, expertise, critical thinking, agency, and responsibility over time.
Human–AI Co-Agency
Adaptive configurations of initiative, expertise, reliance, and responsibility in collaboration.
Longitudinal Human–AI Interaction
Persistent effects of AI-mediated interaction on learning, competence, behavior, and practice.
Advanced Human–AI Evaluation
Psychometric and methodological approaches for richer measures of interaction quality and human outcomes.
Selected contributions
Representative work
Selected publications spanning Interaction Quality, human oversight, responsible Medical AI, and AI in Education.