Behaviour & Information Technology

The evolving human–AI relationship: profiles, trajectories, reciprocal dynamics, and outcomes

Artificial intelligence (AI) is becoming embedded in learning, work, information seeking, decision-making, creativity, communication, and social interaction. Generative AI systems, conversational agents, and intelligent assistants increasingly involve repeated interaction rather than simple technology adoption, with users integrating AI into established practices and assigning it different roles, including those of tool, assistant, collaborator, tutor, advisor, or social agent. Recent research has accordingly highlighted the importance of trust, anthropomorphism, transparency, explainability, verification, and decision delegation in shaping human-AI interaction (Anayat & Kaushik, 2025; Gibbard et al., 2025; Ni & Huang, 2025).

These differences are not adequately captured by frequency or intensity of use alone. Similar levels of AI use may reflect instrumental support, creative collaboration, cognitive offloading, substantial reliance, or delegation of agency. Such patterns may also change with experience, system capabilities, and users' evolving trust and expectations. Recent work further shows that human-like system characteristics, transparency, and explanations can influence trust, reliance, verification, and adoption across different contexts (Gibbard et al., 2025; Gupta & Chauhan, 2026; Ni & Huang, 2025).

Importantly, human-AI interaction is not necessarily a one-way process. Users shape interactions through their prompts, feedback, choices, and interaction histories, while AI-generated responses, recommendations, and explanations may influence users' judgments, emotions, decisions, and subsequent behaviour. Repeated interactions can therefore create feedback loops in which human and AI behaviour affect one another and potentially co-evolve over time. Understanding these processes requires moving beyond approaches that treat AI simply as a technological exposure and users as passive recipients of its effects (Mehrotra et al., 2024).

This Special Issue aims to advance research on profiles, trajectories, and reciprocal dynamics of human-AI interaction. It invites interdisciplinary work examining how people engage with and appropriate AI technologies, what shapes different patterns of interaction, how human and AI behaviours influence one another, how these patterns develop or change over time, and what consequences follow from them. Work examining feedback loops, reciprocal adaptation, interpretability, and, where applicable, the co-evolution of human and AI behaviour is particularly encouraged.

The collection is organised around four closely connected questions:

1. What distinct profiles and trajectories of human-AI interaction and engagement emerge across users, technologies, tasks, and contexts?
2. How are these patterns shaped by users' motivations and characteristics, contextual factors, and the capabilities, affordances, design features, and behaviours of AI technologies?
3. How do human and AI behaviours influence one another across repeated interactions, and under what conditions do these processes lead to reciprocal adaptation, reinforcement, reliance, delegation, or collaboration?
4. How can these interaction patterns and dynamics be interpreted, and how do different profiles and trajectories relate to behavioural, cognitive, social, educational, occupational, wellbeing, and psychosocial outcomes?

Topics of interest include, but are not limited to:

Profiles and patterns of human–AI interaction

• User profiles, interaction typologies, engagement patterns, collaboration and task delegation;
• AI as tool, assistant, collaborator, advisor, tutor, creative partner, or social agent;
• Reliance, verification, cognitive offloading, compensatory use, and overreliance.

Motivations, appropriation, and everyday integration

• Instrumental, educational, occupational, creative, social, emotional, and compensatory motives;
• Psychological needs and gratifications;
• Appropriation and integration of AI into everyday practices and routines.

Trajectories and reciprocal human–AI dynamics

• Longitudinal trajectories and transitions between interaction profiles;
• Development of trust, reliance, habitual use, escalation, reduction, or discontinuation;
• Human–AI feedback loops, reciprocal adaptation, reinforcement, and, where applicable, co-evolution over time.

AI affordances and interaction design

• Personalisation, memory, continuity, anthropomorphism, social presence, and perceived empathy;
• Responsiveness, recommendation, persuasion, and adaptive behaviour;
• Human-centred design, oversight, and human-in-the-loop interaction.

Interpretability, explainability, trust, and agency

• Interpretable and explainable human–AI interaction;
• Understanding why particular profiles, trajectories, and outcomes emerge;
• Transparency, uncertainty, trust calibration, autonomy, agency, verification, and responsible reliance.

Outcomes and individual/contextual differences

• Cognitive, behavioural, educational, occupational, social, and psychosocial outcomes;
• Wellbeing, social support, digital strain, problematic engagement, and dependence;
• Developmental, personality, cultural, contextual, and cross-platform differences.

Methods for studying human–AI interaction

• Person-centred, longitudinal, intensive longitudinal, and experimental designs;
• Interaction logs, digital traces, EMA, sequence and network analysis;
• Interpretable machine learning, computational modelling, qualitative and mixed methods, and measurement development.

Theory and conceptual development

• Conceptual and theoretical models integrating behavioural and AI-centred perspectives;
• Methodological and theoretical contributions advancing the study of human-AI interaction.