University of Zurich

Understanding the Structural Asymmetry in Human-AI Interactions

Talk by Prof. Dr. Andreia Sofia Teixeira

Date: 29.10.26 Time: 12.00 - 13.00 Room: Y27H12

Abstract: Human cognition, mental health and relationships unfold within interconnected social systems that change over time. Across our work, we have studied cognitive-affective structure, the evolution of supportive and conflictual ties in adolescent social networks, and distinct social envirotypes associated with mental health. These perspectives highlight the importance of understanding the human relational ecology in which wellbeing is embedded before asking what changes when conversational AI begins to occupy more social and relational roles. Recent work has argued that human–AI interactions should not be understood only as exchanges between a user and a tool. Conversational systems can simulate responsiveness, validation, emotional language and apparent empathy, and may therefore enter functions that are normally embedded in human relationships. Yet the interaction is structurally asymmetric: the human brings lived experience, vulnerability, embodiment and social consequences, while the system does not participate in the relationship in the same way and does not share reciprocal accountability. In this talk, I will unpack different dimensions of this asymmetry and introduce asymmetric co-regulation as a framework for studying how conversational AI may shape a user's regulatory state without reciprocal regulation in the human sense. I will consider how this asymmetry may interact with anthropomorphism, perceived sociality, memory and personalisation, validation and sycophancy, and the distinction between instrumental and more relational modes of use. I will then ask how repeated interaction may affect regulation and sense of agency over time, when AI may scaffold or substitute for existing human capacities and forms of support, and how effects originating within the human–AI interaction may extend into wider social relationships.

Bio: Andreia Sofia Teixeira is an Associate Professor at the Network Science Institute, Northeastern University London, where she leads the BRAN Lab (Behavioral Research and Adaptive Networks). Her research sits at the intersection of network science and machine learning, developing measures, computational models and simulation frameworks to understand the structure and dynamics of complex systems, and how local behaviour can spill over to collective scales. Her work spans human–AI interaction, collective behaviour, cognitive and affective networks, signed and temporal social networks, mental health and neuropsychiatric and neurodegenerative disorders. Across these areas, she is particularly interested in how individual states, relationships and broader social environments shape one another over time.