From Spectator to Co-Creator: A Social XR Framework with Generative AI

Immersive attractions promise shared wonder, yet many still treat audiences as passive consumers of media. My doctoral research reframes Extended Reality (XR) as a socially co-constructed space whose success depends on how groups negotiate norms with each other and with Generative AI agents. In partnership with Holovis International, I build room-scale immersive experiences where GenAI characters, audiovisuals, interactions adapt in real time to the dynamics of the group. Through Convergent Mixed-Methods – observation of behaviour, semi-structured interviews and live player analytics, I seek to reveal which design levers cultivate pro social cooperation and the moment spectators become co-creators of the experience. The resulting Social XR Framework translates my research into evidence-based and industry-backed guidelines for immersive entertainment designers, enabling scalable, personalised experiences that strengthen human connection in the age of generative AI

Keywords:

Transdisciplinary, Frameworks, Interactive, Immersive, Mixed Reality, Generative AI, Real-time Data

Key Research Areas for Transdisciplinary Focus: Human Computing Interaction (HCI) and Emerging Multimedia Technologies, Game Studies, Social Computing (Norm Dynamics).

Key Industry Areas:

Themed Entertainment – Immersive Production – Room-Scale / Co-Located XR Attractions

With adjacent applications to Immersive Retail (due to Theme Parks being hybrid spaces of entertainment and retail)

Methodology and Worldview

Practice-Based Research informed by Mixed Methods Analysis (PBR + Interviews + Observation Data)

Constructivist Worldview

Research Questions and Methodology

The literature reviews I have conducted allowed me to deeply explore XR frameworks, co-creation, immersion, and social dynamics, all core concepts that have revealed critical gaps. The existing scholarship often focuses on technology or individual psychology, neglecting the socio-technical connection, particularly:

(1) Limited understanding of how AI-Adaptive Co-Located XR influences the Norm Formation of Participants

(2) Treatment of AI as content rather than as a social element in XR

(3 )Lack of comparative evidence on interaction or immersive configurations that foster pro-social cooperation from the participant’s perspective.

(4) Limited understanding of how participants interpret and react to AI Behaviour during norm negotiations or violations

The gaps outlined above have thus informed me on the further development of my four research questions (RQs):

RQ1: How do participants perceive and interpret social norms within generative-AI-mediated social-XR experiences, and how do those perceptions evolve?

RQ2: How do participants interpret and respond to the AI agent’s adaptive behaviour, and how does this influence the emergence, reinforcement, or violation of socially constructed norms?

RQ3:From the participants’ perspective and observed interactions, what configurations of system design features, participant attributes, and group context are associated with cooperative norms?

RQ4: How do participants respond to perceived violations of social norms by other participants or resulting from the AI agent’s behaviour?

To address the research questions above, I require a research paradigm that enables me to understand phenomena that are fundamentally subjective (such as interpretations and perception), interactional (behaviors within a social context) and emergent (group dynamics based on shared meaning and norms). As such, I have found and selected the constructivist approach as it is ideal in helping me frame and capture the multiple, socially constructed realities and processes through which they develop (Guba and Lincoln, 1994). To bridge the literature gaps highlighted, I find the need for methods that can explore subjective meaning and interaction patterns within XR experience. I have adopted a Convergent Mixed-Methods design (Creswell and Plano Clark, 2018), to allow for the integration of qualitative data, essential for understanding participant perspectives and constructed meanings, with quantitative data, of which provides contextual information about behaviours and systems interactions.

In this research, I operationalise the mixed-methods design through a Practice-Based Research (PBR) strategy (Candy and Edmonds, 2018; Vear, 2018). PBR positions the novel prototype developed through practice as an artifact, of which forms the basis for generating situated understanding and contribution to knowledge. As Candy and Edmonds (2018, p.63) state, “practice-based research is an original investigation undertaken in order to gain new knowledge, partly by means of practice and the outcomes of that practice”, to which my background as a video game developer, and a creative technologist specializing in emerging technology affords me the ability to truly develop the creative practice. Through iterative design, the creative practice continually deploys and refines the prototypes and thus creates a loop that informs both the next prototype iteration and deepening my understanding necessary to answer the RQs.