The Psychology of Collaborating and Co-Creating with Artificial Intelligence

This dissertation, structured around three chapters, explores the nexus between employee creativity, collaboration, and artificial intelligence (AI) in organizational settings. The broader motivation for this research stems from the growing presence of AI in knowledge and creative work, which raises fundamental theoretical and practical questions about how human capabilities are redefined in relation to intelligent AI systems. As organizations increasingly adopt AI tools not just for automation or augmentation but for ideation and collaboration, it becomes essential to understand how employees make sense of, adapt to, and co-create with AI. This thesis examines long-standing research questions about employee creativity and collaboration and investigates AI’s impact on employee behavior. To build theory on this timely topic, I employed quantitative and qualitative methods to bring a well-rounded approach to our understanding of the psychological and behavioral implications of collaborating and co-creating with AI. This dissertation includes a qualitative investigation of fashion designers and stylists co-creating with AI (Paper 1), a theoretical exploration of creative autonomy when working with generative AI (Paper 2), and a mixed-methods examination of social comparisons and idea integration with generative AI (Paper 3). Although these papers are independent studies, they have informed and shaped each other, contributing to an overall narrative on human-AI co-creation. To deepen our theoretical understanding of creating with AI, I took an inductive approach and conducted a nearly three-year longitudinal case study with fashion designers and stylists who work directly with AI suggestions. This study was driven by my curiosity about how creative professionals interact and work with AI when they are required to simultaneously produce creative work and incorporate constraining, often opaque, AI suggestions. The first paper, “Stitching Together: Employee-AI Creative Collaboration in Fashion Design and Styling,” examines this puzzle and builds theory on collaboration and creativity . While collaboration is a cornerstone of creative work, it is unclear how conventional paradigms of working in dyads, groups, and teams translate to situations where individuals are required to work with AI rather than human colleagues, with whom mutual communication, information exchange, and shared objectives are common (Amabile et al., 2001). This paper untangles this intricate process and unveils the cognitive and psychological mechanisms that underlie how and why employees incorporate AI suggestions into their work. The findings illustrate the dynamic ways fashion creatives initially lean in to learn about co-working with AI, disconnect to preserve their creative freedom and creative identity, and eventually reconcile ways to reconnect with AI for co-existence. In the second paper, “Creative Autonomy in the Age of Generative AI: A Typology of Human-AI Creative Interaction,” I explore the complex interplay between creative autonomy and generative AI in creative work, proposing a novel theoretical framework that reconceptualizes autonomy as dynamically enacted through human-AI interactions. As a first step, I introduce two key constructs: ideational steering, which refers to how generative AI influences creative thinking, and experienced creative agency, which is the individual’s sense of control over their creative process. Next, the framework identifies four distinct modes of interaction—autonomous exploration, co-creation, anchored compliance, and creative drift—each reflecting varying levels of AI influence and individual agency. This typology serves to clarify how AI can simultaneously enhance and constrain creativity, depending on individuals’ engagement with its ideas, suggestions, or outputs. By doing so, it shifts the focus of human-AI research from the functional use of AI to the subjective experience of interaction, offers a process-based account of creative autonomy, and identifies modes of creative misalignment. Together, these insights provide a theoretical foundation for understanding how creative autonomy is enacted, sustained, or diminished with generative AI. The third paper, “Co-Creating with AI: How Social Comparisons with AI Affect Idea Integration,” adopts a social comparison framework to examine whether individuals engage in social comparison processes with AI and the extent to which this affects idea integration with generative AI . Social comparison theory (Festinger, 1954) postulates that individuals have an innate drive to evaluate their abilities and opinions by comparing themselves to others. Historically, this theory has been applied to human-human interactions. The notion that individuals will engage in social comparison with machines may initially seem unconventional, but given that the fundamental purpose of AI is to mimic human abilities (McCarthy, Minsky, Rochester, & Shannon, 1955) and replace jobs (e.g., Ferràs-Hernández, 2018), this likely compels individuals to view AI as a comparison target. Employing a mixed methods approach, we conducted two studies to investigate the psychological underpinnings that influence individuals’ decisions to accept, reject, or integrate AI-generated ideas. This paper shows that the fundamental human inclination to evaluate the self in relation to others extends to AI as well. Participants interacting with an AI expert engaged in more idea integration when they were led to believe they are similar in abilities to a co-partner (lateral) than one that is inferior (downward) or superior (upward). Additionally, we found that cognitive trust plays an important role. Our thematic analysis revealed that participants actively seek ways to avoid extremely positive or negative sentiments about AI for a balanced perspective, they take on dynamic roles as a student, guide, or collaborator when working with AI, and they hold layered perceptions of AI’s micro benefits and macro risks. Taken together, these three papers revealed how humans co-create with AI in organizational settings. We found throughout that interactions with AI are driven by psychological processes, and these processes result in different behavioral responses. By extending foundational theories of creativity, autonomy, collaboration, and social comparison into the area of human-AI interaction, this dissertation contributes a novel, empirically grounded understanding of how individuals psychologically and behaviorally navigate co-creation with AI. More broadly, it advances organizational scholarship by introducing process-oriented, experience-based frameworks that account for the unique dynamics of working with generative AI. Doing so highlights that the future of creative work will be largely determined by how humans interpret, resist, and collaborate with these advanced generative AI systems.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2024

Author dc:creator

• Vardanyan, Alentina

Advisor dc:contributor.advisor

• Richter, Andreas

• creativity
• artificial intelligence
• human-AI collaboration
• human-AI co-creation
• social comparison
• collaboration

Licence

• https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/a113a3be-fea3-420f-8d69-5b509c042883/download
• http://purl.org/NET/rdflicense/allrightsreserved

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.120229

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/387461