Large language models (LLMs) are popular tools for creative ideation, but have been shown to homogenize outputs across users. We test if approaching an AI tool with high human (vs. low) human agency can mitigate this homogenization effect by encouraging people to use AI to augment their creativity, rather than offload it. Participants were experimentally assigned to one of three conditions (high-agency approach with AI access, low-agency approach with AI access, or a human-only control) and generated creative uses for everyday objects. Contrary to our expectations, ideas did not differ in individual-level quality (overall creativity, originality, and usefulness). However, a preregistered similarity-to-centroid analysis and an exploratory cluster analysis provided convergent evidence of AI-induced homogenization among the low-agency condition. Thus, while AI has enabled greater ideational fluency, our research suggests the degree of agency with which people approach their AI tool has downstream consequences on collective creativity.

Concerns about smartphone dependency have sparked interest in minimal mobile phones: devices supporting basic communication without social apps, web browsing, or games. These design choices are thought to improve well-being, but have not been tested empirically. We conducted a first-of-its-kind longitudinal experiment examining the effects of switching from smartphones to minimal mobile phones on young adults’ psychological well-being over a week (n = 166). To account for individual variation in intrinsic motivation to try a minimal phone, we employed a quasi-experimental design comparing the outcomes of three groups: 1) high-interest volunteers who were asked to use minimal phones or participants that were randomly assigned to either 2) use minimal phones or 3) continue using their own smartphones. Results showed that switching to a minimal phone effectively reduced phone and social media use. However, only high-interest volunteers - those intrinsically motivated to participate - showed significant within-person changes in psychological well-being, reporting reduced stress, increased life satisfaction, and less FoMo. No effects on well-being were observed for those assigned to use the phone. Our results suggest that switching to a minimal mobile phone may support some motivated individuals in improving their sense of agency and well-being in an increasingly connected digital world.

Personalized algorithms are central to how people discover information and engage with media online. Drawing on interviews and screen-sharing sessions with TikTok users (N=27), we extend the algorithmic crystal framework, which conceptualizes personalized algorithms as reflective surfaces through which users may interpret their experiences with content in relation to their own self-concepts. This research expands the framework to account for the interpersonal dynamics that emerge from user engagement with algorithmic feeds. We found that users who feel ''seen'' by the algorithm use its personalized content recommendations for social signaling: sharing content that represents themselves (''this is me''), acknowledges how they see others (''this is you''), and affirms shared identities (''this is us''). We suggest that these dynamics give rise to a hybrid form of digital selfhood simultaneously shaped by algorithmic profiling and networked social interaction-blurring existing separations in digital identity theory. We also build on the concept of diffracted belonging-the experience of recognizing aspects of oneself in the content of diverse others-to explore how users interpret algorithmically-recommended content as reflective of the self. Our findings suggest that such moments of recognition may contribute to shifts in self-perception and support ongoing processes of identity development. Finally, we illustrate how users engage in the strategic refinement of their feeds to manage how they feel while using the platform. Our findings suggest that this process involves reflective, and sometimes effortful, negotiation with the algorithm, highlighting the co-produced nature of mood management in everyday human-algorithm interactions. Together, these findings underscore the interpersonal and psychological dynamics of interacting with personalized algorithms and provide insights into how social communication and identity work unfold in algorithmically-mediated environments.

This research examines how TikTok users conceptualize and engage with personalized algorithms on the TikTok platform. Using qualitative methods, we analyzed 24 interviews with TikTok users to explore how algorithmic personalization processes inform people's understanding of their identities as well as shape their orientation to others. Building on insights from our qualitative data and previous scholarship on algorithms and identity, we propose a novel conceptual model to understand how people think about and interact with personalized algorithmic systems. Drawing on the metaphor of crystals and their properties, the algorithmic crystal framework is an analytic frame that captures user understandings of how personalized algorithms (1) interact with user identity by reflecting user self-concepts that are both multifaceted and dynamic and (2) shape perspectives on others encountered through the algorithm, by orienting users to recognize parts of themselves refracted in other users and to experience ephemeral, diffracted connections with groups of similar others. We describe how the algorithmic crystal framework can extend theory and inform new lines of research around the implications of algorithms in self-concept development and social life.